_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d16299141 | Recent neural network sequence models with softmax classifiers have achieved their best language modeling performance only with very large hidden states and large vocabularies. Even then they struggle to predict rare or unseen words even if the context makes the prediction unambiguous. We introduce the pointer sentinel mixture architecture for neural sequence models which has the ability to either reproduce a word from the recent context or produce a word from a standard softmax classifier. Our pointer sentinel-LSTM model achieves state of the art language modeling performance on the Penn Treebank (70.9 perplexity) while using far fewer parameters than a standard softmax LSTM. In order to evaluate how well language models can exploit longer contexts and deal with more realistic vocabularies and larger corpora we also introduce the freely available WikiText corpus. 1 | Pointer Sentinel Mixture Models |
d252735240 | Meta-training, which fine-tunes the language model (LM) on various downstream tasks by maximizing the likelihood of the target label given the task instruction and input instance, has improved the zero-shot task generalization performance. However, meta-trained LMs still struggle to generalize to challenging tasks containing novel labels unseen during meta-training. In this paper, we propose FLIPPED LEARNING, an alternative method of meta-training which trains the LM to generate the task instruction given the input instance and label. During inference, the LM trained with FLIPPED LEARNING, referred to as FLIPPED, selects the label option that is most likely to generate the task instruction. On 14 tasks of the BIG-bench benchmark, the 11B-sized FLIPPED outperforms zero-shot T0-11B (Sanh et al., 2021) and even a 16 times larger 3-shot GPT-3 (175B) (Brown et al., 2020) on average by 8.4% and 9.7% points, respectively. FLIPPED gives particularly large improvements on tasks with unseen labels, outperforming T0-11B by up to +20% average F1 score. This indicates that the strong task generalization of FLIPPED comes from improved generalization to novel labels. We release our code at github.com/seonghyeonye/Flipped-Learning. | GUESS THE INSTRUCTION! FLIPPED LEARNING MAKES LANGUAGE MODELS STRONGER ZERO-SHOT LEARNERS |
d252595881 | This work introduces DiGress, a discrete denoising diffusion model for generating graphs with categorical node and edge attributes. Our model utilizes a discrete diffusion process that progressively edits graphs with noise, through the process of adding or removing edges and changing the categories. A graph transformer network is trained to revert this process, simplifying the problem of distribution learning over graphs into a sequence of node and edge classification tasks. We further improve sample quality by introducing a Markovian noise model that preserves the marginal distribution of node and edge types during diffusion, and by incorporating auxiliary graph-theoretic features. A procedure for conditioning the generation on graph-level features is also proposed. DiGress achieves state-of-theart performance on molecular and non-molecular datasets, with up to 3x validity improvement on a planar graph dataset. It is also the first model to scale to the large GuacaMol dataset containing 1.3M drug-like molecules without the use of molecule-specific representations. * Equal contribution. Contact: first name.last name@epfl.ch arXiv:2209.14734v4 [cs.LG] 23 May 2023Published as a conference paper at ICLR 2023 T t=2 q(z t |z t−1 ). The denoising network ϕ θ is trained to invert this process by predicting z t−1 from z t . To generate new samples, noise is sampled from a prior distribution and then inverted by iterative application of the denoising network. | DIGRESS: DISCRETE DENOISING DIFFUSION FOR GRAPH GENERATION |
d17299045 | Network quantization is one of network compression techniques employed to reduce the redundancy of deep neural networks. It compresses the size of the storage for a large number of network parameters in a neural network by quantizing them and encoding the quantized values into binary codewords of smaller sizes. In this paper, we aim to design network quantization schemes that minimize the expected loss due to quantization while maximizing the compression ratio. To this end, we analyze the quantitative relation of quantization errors to the loss function of a neural network and identify that the Hessian-weighted distortion measure is locally the right objective function that we need to optimize for minimizing the loss due to quantization. As a result, Hessian-weighted k-means clustering is proposed for clustering network parameters to quantize when fixed-length binary encoding follows. Hessian-weighting properly handles the different impact of quantization errors not only within layers but also across layers and thus it can be employed for quantizing all layers of a network together at once; it is beneficial since one can avoid layer-by-layer compression rate optimization. When optimal variablelength binary codes, e.g., Huffman codes, are employed for further compression of quantized values after clustering, we derive that the network quantization problem can be related to the entropy-constrained scalar quantization (ECSQ) problem in information theory and consequently propose two solutions of ECSQ for network quantization, i.e., uniform quantization and an iterative algorithm similar to Lloyd's algorithm for k-means clustering. Finally, using the simple uniform quantization followed by Huffman coding, our experiment results show that the compression ratios of 51.25, 22.17 and 40.65 are achievable (i.e., the sizes of the compressed models are 1.95%, 4.51% and 2.46% of the original model sizes) for LeNet, ResNet and AlexNet, respectively, at no or marginal performance loss.The large sizes of deep neural networks make it difficult to deploy them on resource-limited devices, e.g., mobile or embedded devices. However, it has been shown that deep neural networks typically include many redundant parameters(Denil et al., 2013)and it is of great interest in recent years to reduce the redundancy of deep neural networks.The whole procedure to reduce the redundancy of a neural network is called as network compression. The benefit of compression is twofold: computational cost reduction and hardware resource saving. In this paper, our interest is mainly on the size reduction of the storage (memory) for a large number of network parameters (weights and biases). In particular, we focus on the network size compression by quantizing network parameters. The most related work to our investigation can be found inGong et al. (2014);Han et al. (2015a), where a conventional quantization method | TOWARDS THE LIMIT OF NETWORK QUANTIZATION |
d10135357 | Recurrent Neural Networks (RNN) are widely used to solve a variety of problems and as the quantity of data and the amount of available compute have increased, so have model sizes. The number of parameters in recent state-of-the-art networks makes them hard to deploy, especially on mobile phones and embedded devices. The challenge is due to both the size of the model and the time it takes to evaluate it. In order to deploy these RNNs efficiently, we propose a technique to reduce the parameters of a network by pruning weights during the initial training of the network. At the end of training, the parameters of the network are sparse while accuracy is still close to the original dense neural network. The network size is reduced by 8× and the time required to train the model remains constant. Additionally, we can prune a larger dense network to achieve better than baseline performance while still reducing the total number of parameters significantly. Pruning RNNs reduces the size of the model and can also help achieve significant inference time speed-up using sparse matrix multiply. Benchmarks show that using our technique model size can be reduced by 90% and speed-up is around 2× to 7×.Published as a conference paper at ICLR 2017The more powerful server class GPUs used in data centers can generally perform inference quickly enough to serve one user, but in the data center performance per dollar is very important. Techniques that allow models to be evaluated faster enable more users to be served per GPU increasing the effective performance per dollar. | EXPLORING SPARSITY IN RECURRENT NEURAL NETWORKS |
d235446639 | We consider the problem of estimating the number of distinct elements in a large data set (or, equivalently, the support size of the distribution induced by the data set) from a random sample of its elements. The problem occurs in many applications, including biology, genomics, computer systems and linguistics. A line of research spanning the last decade resulted in algorithms that estimate the support up to ±εn from a sample of size O(log 2 (1/ε) · n/ log n), where n is the data set size. Unfortunately, this bound is known to be tight, limiting further improvements to the complexity of this problem. In this paper we consider estimation algorithms augmented with a machine-learning-based predictor that, given any element, returns an estimation of its frequency. We show that if the predictor is correct up to a constant approximation factor, then the sample complexity can be reduced significantly, to log(1/ε) · n 1−Θ(1/ log(1/ε)) .We evaluate the proposed algorithms on a collection of data sets, using the neuralnetwork based estimators from Hsu et al, ICLR'19 as predictors. Our experiments demonstrate substantial (up to 3x) improvements in the estimation accuracy compared to the state of the art algorithm. | LEARNING-BASED SUPPORT ESTIMATION IN SUBLINEAR TIME |
d3986339 | Real-valued word representations have transformed NLP applications; popular examples are word2vec and GloVe, recognized for their ability to capture linguistic regularities. In this paper, we demonstrate a very simple, and yet counterintuitive, postprocessing technique -eliminate the common mean vector and a few top dominating directions from the word vectors -that renders off-the-shelf representations even stronger. The postprocessing is empirically validated on a variety of lexical-level intrinsic tasks (word similarity, concept categorization, word analogy) and sentence-level extrinsic tasks (semantic textual similarity) on multiple datasets and with a variety of representation methods and hyperparameter choices in multiple languages; in each case, the processed representations are consistently better than the original ones. Furthermore, we demonstrate quantitatively in downstream applications that neural network architectures "automatically learn" the postprocessing operation. . 2013. Translating embeddings for modeling multirelational data. In Advances in Neural Information Processing Systems. pages 2787-2795. Michael W Browne. 1979. The maximum-likelihood solution in inter-battery factor analysis. British Journal of Mathematical and Statistical Psychology 32(1):75-86. | All-but-the-Top: Simple and Effective Postprocessing for Word Representations |
d202734152 | Cloud computing is becoming increasingly popular as a platform for distributed training of deep neural networks. | GAP-AWARE MITIGATION OF GRADIENT STALENESS |
d257636846 | Continual learning is a problem for artificial neural networks that their biological counterparts are adept at solving. Building on work using Sparse Distributed Memory (SDM) to connect a core neural circuit with the powerful Transformer model, we create a modified Multi-Layered Perceptron (MLP) that is a strong continual learner. We find that every component of our MLP variant translated from biology is necessary for continual learning. Our solution is also free from any memory replay or task information, and introduces novel methods to train sparse networks that may be broadly applicable.Published as a conference paper at ICLR 2023 in the class incremental setting when memory replay is not allowed. Another variant of SDM, developed independently of our work, appears to be state-of-the-art for CIFAR-100, MNIST, and FashionMNIST, with our SDMLP as a close second(Shen et al., 2021).Excitingly, our continual learning success is "organic" in resulting from the underlying model architecture and does not require any task labels, task boundaries, or memory replay. Abstractly, SDM learns its subnetworks responsible for continual learning using two core model components. First, the Top-K activation function causes the k neurons most activated by an input to specialize towards this input, resulting in the formation of specialized subnetworks. Second, the L 2 normalization and absence of a bias term together constrain neurons to the data manifold, ensuring that all neurons democratically participate in learning. When these two components are combined, a new learning task only activates and trains a small subset of neurons, leaving the rest of the network intact to remember previous tasks without being overwritten As a roadmap of the paper, we first discuss related work (Section 2). We then provide a short introduction to SDM (Section 3), before translating it into our MLP (Section 4). Next we present our results, comparing the organic continual learning capabilities of our SDMLP against relevant benchmarks (Section 5). Finally, we conclude with a discussion on the limitations of our work, sparse models more broadly, and how SDM relates MLPs to Transformer Attention (Section 6). | SPARSE DISTRIBUTED MEMORY IS A CONTINUAL LEARNER |
d235829401 | Most existing Vision-and-Language (V&L) models rely on pre-trained visual encoders, using a relatively small set of manuallyannotated data (as compared to web-crawled data), to perceive the visual world. However, it has been observed that large-scale pretraining usually can result in better generalization performance, e.g., CLIP (Contrastive Language-Image Pre-training), trained on a massive amount of image-caption pairs, has shown a strong zero-shot capability on various vision tasks. To further study the advantage brought by CLIP, we propose to use CLIP as the visual encoder in various V&L models in two typical scenarios: 1) plugging CLIP into task-specific fine-tuning; 2) combining CLIP with V&L pre-training and transferring to downstream tasks. We show that CLIP significantly outperforms widely-used visual encoders trained with in-domain annotated data, such as BottomUp-TopDown. We achieve competitive or better results on diverse V&L tasks, while establishing new state-of-the-art results on Visual Question Answering, Visual Entailment, and V&L Navigation tasks.1 Without confusion, we use the term CLIP to interchangeably refer to both the whole CLIP model (including the text and visual encoder) and just its visual encoder. 2 We distinguish between V&L pre-training and CLIP pretraining: V&L pre-training models have deep interactions between modalities while CLIP follows a shallow-interaction design (Section 2). | How Much Can CLIP Benefit Vision-and-Language Tasks? |
d247518670 | We propose a new method for spatio-temporal forecasting on arbitrarily distributed points. Assuming that the observed system follows an unknown partial differential equation, we derive a continuous-time model for the dynamics of the data via the finite element method. The resulting graph neural network estimates the instantaneous effects of the unknown dynamics on each cell in a meshing of the spatial domain. Our model can incorporate prior knowledge via assumptions on the form of the unknown PDE, which induce a structural bias towards learning specific processes. Through this mechanism, we derive a transport variant of our model from the convection equation and show that it improves the transfer performance to higher-resolution meshes on sea surface temperature and gas flow forecasting against baseline models representing a selection of spatio-temporal forecasting methods. A qualitative analysis shows that our model disentangles the data dynamics into their constituent parts, which makes it uniquely interpretable. | LEARNING THE DYNAMICS OF PHYSICAL SYSTEMS FROM SPARSE OBSERVATIONS WITH FINITE ELEMENT NETWORKS |
d253264964 | Enabling private inference is crucial for many cloud inference services that are based on Transformer models. However, existing private inference solutions can increase the inference latency by more than 60× or significantly compromise the inference quality. In this paper, we design the framework MPCFORMER as a practical solution, using Secure Multi-Party Computation (MPC) and Knowledge Distillation (KD). Through extensive evaluations, we show that MPCFORMER significantly speeds up Transformer inference in MPC settings while achieving similar ML performance to the input model. On the IMDb dataset, it achieves similar performance to BERT BASE , while being 5.3× faster. On the GLUE benchmark, it achieves 97% performance of BERT BASE with a 2.2× speedup. MPC-FORMER remains effective with different trained Transformer weights such as ROBERTA BASE and larger models including BERT Large . Code is available at https://github.com/MccRee177/MPCFormer. * Authors contributed equally. 1 We will use the term MPC-friendly approximations. | MPCFORMER: FAST, PERFORMANT AND PRIVATE TRANSFORMER INFERENCE WITH MPC |
d21731691 | Explicit encoding of group actions in deep features makes it possible for convolutional neural networks (CNNs) to handle global deformations of images, which is critical to success in many vision tasks. This paper proposes to decompose the convolutional filters over joint steerable bases across the space and the group geometry simultaneously, namely a rotationequivariant CNN with decomposed convolutional filters (RotDCF). This decomposition facilitates computing the joint convolution, which is proved to be necessary for the group equivariance. It significantly reduces the model size and computational complexity while preserving performance, and truncation of the bases expansion serves implicitly to regularize the filters. On datasets involving in-plane and out-of-plane object rotations, RotDCF deep features demonstrate greater robustness and interpretability than regular CNNs. The stability of the equivariant representation to input variations is also proved theoretically under generic assumptions on the filters in the decomposed form. The RotDCF framework can be extended to groups other than rotations, providing a general approach which achieves both group equivariance and representation stability at a reduced model size. * | RotDCF: Decomposition of Convolutional Filters for Rotation-Equivariant Deep Networks |
d254823652 | The robustness of machine learning algorithms to distributions shift is primarily discussed in the context of supervised learning (SL). As such, there is a lack of insight on the robustness of the representations learned from unsupervised methods, such as self-supervised learning (SSL) and auto-encoder based algorithms (AE), to distribution shift. We posit that the input-driven objectives of unsupervised algorithms lead to representations that are more robust to distribution shift than the target-driven objective of SL. We verify this by extensively evaluating the performance of SSL and AE on both synthetic and realistic distribution shift datasets. Following observations that the linear layer used for classification itself can be susceptible to spurious correlations, we evaluate the representations using a linear head trained on a small amount of out-of-distribution (OOD) data, to isolate the robustness of the learned representations from that of the linear head. We also develop "controllable" versions of existing realistic domain generalisation datasets with adjustable degrees of distribution shifts. This allows us to study the robustness of different learning algorithms under versatile yet realistic distribution shift conditions. Our experiments show that representations learned from unsupervised learning algorithms generalise better than SL under a wide variety of extreme as well as realistic distribution shifts. | HOW ROBUST IS UNSUPERVISED REPRESENTATION LEARNING TO DISTRIBUTION SHIFT? |
d221112371 | To get Bayesian neural networks to perform comparably to standard neural networks it is usually necessary to artificially reduce uncertainty using a "tempered" or "cold" posterior. This is extremely concerning: if the prior is accurate, Bayes inference/decision theory is optimal, and any artificial changes to the posterior should harm performance. While this suggests that the prior may be at fault, here we argue that in fact, BNNs for image classification use the wrong likelihood. In particular, standard image benchmark datasets such as CIFAR-10 are carefully curated. We develop a generative model describing curation which gives a principled Bayesian account of cold posteriors, because the likelihood under this new generative model closely matches the tempered likelihoods used in past work. | A statistical theory of cold posteriors in deep neural networks |
d256416103 | Recent state-of-the-art source-free domain adaptation (SFDA) methods have focused on learning meaningful cluster structures in the feature space, which have succeeded in adapting the knowledge from source domain to unlabeled target domain without accessing the private source data. However, existing methods rely on the pseudo-labels generated by source models that can be noisy due to domain shift. In this paper, we study SFDA from the perspective of learning with label noise (LLN). Unlike the label noise in the conventional LLN scenario, we prove that the label noise in SFDA follows a different distribution assumption. We also prove that such a difference makes existing LLN methods that rely on their distribution assumptions unable to address the label noise in SFDA. Empirical evidence suggests that only marginal improvements are achieved when applying the existing LLN methods to solve the SFDA problem. On the other hand, although there exists a fundamental difference between the label noise in the two scenarios, we demonstrate theoretically that the early-time training phenomenon (ETP), which has been previously observed in conventional label noise settings, can also be observed in the SFDA problem. Extensive experiments demonstrate significant improvements to existing SFDA algorithms by leveraging ETP to address the label noise in SFDA. . Learning with instancedependent label noise: A sample sieve approach. arXiv preprint arXiv: | WHEN SOURCE-FREE DOMAIN ADAPTATION MEETS LEARNING WITH NOISY LABELS |
d246822636 | One critical component in lossy deep image compression is the entropy model, which predicts the probability distribution of the quantized latent representation in the encoding and decoding modules. Previous works build entropy models upon convolutional neural networks which are inefficient in capturing global dependencies. In this work, we propose a novel transformer-based entropy model, termed Entroformer, to capture long-range dependencies in probability distribution estimation effectively and efficiently. Different from vision transformers in image classification, the Entroformer is highly optimized for image compression, including a top-k self-attention and a diamond relative position encoding. Meanwhile, we further expand this architecture with a parallel bidirectional context model to speed up the decoding process. The experiments show that the Entroformer achieves state-of-the-art performance on image compression while being time-efficient. Code is available at https://github.com/damo-cv/entroformer. | ENTROFORMER: A TRANSFORMER-BASED ENTROPY MODEL FOR LEARNED IMAGE COMPRESSION |
d246652106 | Face clustering has attracted rising research interest recently to take advantage of massive amounts of face images on the web. State-of-the-art performance has been achieved by Graph Convolutional Networks (GCN) due to their powerful representation capacity. However, existing GCN-based methods build face graphs mainly according to kNN relations in the feature space, which may lead to a lot of noise edges connecting two faces of different classes. The face features will be polluted when messages pass along these noise edges, thus degrading the performance of GCNs. In this paper, a novel algorithm named Ada-NETS is proposed to cluster faces by constructing clean graphs for GCNs. In Ada-NETS, each face is transformed to a new structure space, obtaining robust features by considering face features of the neighbour images. Then, an adaptive neighbour discovery strategy is proposed to determine a proper number of edges connecting to each face image. It significantly reduces the noise edges while maintaining the good ones to build a graph with clean yet rich edges for GCNs to cluster faces. Experiments on multiple public clustering datasets show that Ada-NETS significantly outperforms current state-of-the-art methods, proving its superiority and generalization. Code is available at https://github.com/damo-cv/Ada-NETS. | ADA-NETS: FACE CLUSTERING VIA ADAPTIVE NEIGHBOUR DISCOVERY IN THE STRUCTURE SPACE |
d252596252 | Diffusion models have been recently studied as powerful generative inverse problem solvers, owing to their high quality reconstructions and the ease of combining existing iterative solvers. However, most works focus on solving simple linear inverse problems in noiseless settings, which significantly under-represents the complexity of real-world problems. In this work, we extend diffusion solvers to efficiently handle general noisy (non)linear inverse problems via approximation of the posterior sampling. Interestingly, the resulting posterior sampling scheme is a blended version of diffusion sampling with the manifold constrained gradient without a strict measurement consistency projection step, yielding a more desirable generative path in noisy settings compared to the previous studies. Our method demonstrates that diffusion models can incorporate various measurement noise statistics such as Gaussian and Poisson, and also efficiently handle noisy nonlinear inverse problems such as Fourier phase retrieval and non-uniform deblurring. | DIFFUSION POSTERIOR SAMPLING FOR GENERAL NOISY INVERSE PROBLEMS |
d198986015 | Generative models are emerging as promising tools in robotics and reinforcement learning. Yet, even though tasks in these domains typically involve distinct objects, most state-of-the-art methods do not explicitly capture the compositional nature of visual scenes. Two exceptions, MONet and IODINE, decompose scenes into objects in an unsupervised fashion via a set of latent variables. Their underlying generative processes, however, do not account for component interactions. Hence, neither of them allows for principled sampling of coherent scenes. Here we present GENESIS, the first object-centric generative model of visual scenes capable of both decomposing and generating complete scenes by explicitly capturing relationships between scene components. GENESIS parameterises a spatial GMM over pixels which is encoded by component-wise latent variables that are inferred sequentially or sampled from an autoregressive prior. We train GENESIS on two publicly available datasets and probe the information in the latent representations through a set of classification tasks, outperforming several baselines. | GENESIS: Generative Scene Inference and Sampling with Object-Centric Latent Representations |
d235436157 | Domain adaptation is an important problem and often needed for real-world applications. In this problem, instead of i.i.d. training and testing datapoints, we assume that the source (training) data and the target (testing) data have different distributions. With that setting, the empirical risk minimization training procedure often does not perform well, since it does not account for the change in the distribution. A common approach in the domain adaptation literature is to learn a representation of the input that has the same (marginal) distribution over the source and the target domain. However, these approaches often require additional networks and/or optimizing an adversarial (minimax) objective, which can be very expensive or unstable in practice. To improve upon these marginal alignment techniques, in this paper, we first derive a generalization bound for the target loss based on the training loss and the reverse Kullback-Leibler (KL) divergence between the source and the target representation distributions. Based on this bound, we derive an algorithm that minimizes the KL term to obtain a better generalization to the target domain. We show that with a probabilistic representation network, the KL term can be estimated efficiently via minibatch samples without any additional network or a minimax objective. This leads to a theoretically sound alignment method which is also very efficient and stable in practice. Experimental results also suggest that our method outperforms other representation-alignment approaches. | KL GUIDED DOMAIN ADAPTATION |
d258833272 | The online knapsack problem is a classic problem in the field of online algorithms. Its canonical version asks how to pack items of different values and weights arriving online into a capacity-limited knapsack so as to maximize the total value of the admitted items. Although optimal competitive algorithms are known for this problem, they may be fundamentally unfair, i.e., individual items may be treated inequitably in different ways. Inspired by recent attention to fairness in online settings, we develop a natural and practically-relevant notion of time fairness for the online knapsack problem, and show that the existing optimal algorithms perform poorly under this metric. We propose a parameterized deterministic algorithm where the parameter precisely captures the Pareto-optimal trade-off between fairness and competitiveness. We show that randomization is theoretically powerful enough to be simultaneously competitive and fair; however, it does not work well in practice, using trace-driven experiments. To further improve the trade-off between fairness and competitiveness, we develop a fair, robust (competitive), and consistent learning-augmented algorithm with substantial performance improvement in trace-driven experiments.Preprint. Under review. | Time Fairness in Online Knapsack Problems |
d239049848 | Reinforcement learning (RL) experiments have notoriously high variance, and minor details can have disproportionately large effects on measured outcomes.This is problematic for creating reproducible research and also serves as an obstacle when applying RL to sensitive real-world applications.In this paper, we investigate causes for this perceived instability.To allow for an in-depth analysis, we focus on a specifically popular setup with high variance -continuous control from pixels with an actor-critic agent.In this setting, we demonstrate that poor outlier runs which completely fail to learn are an important source of variance, but that weight initialization and initial exploration are not at fault.We show that one cause for these outliers is unstable network parametrization which leads to saturating nonlinearities.We investigate several fixes to this issue and find that simply normalizing penultimate features is surprisingly effective.For sparse tasks, we also find that partially disabling clipped double Q-learning decreases variance.By combining fixes we significantly decrease variances, lowering the average standard deviation across 21 tasks by a factor > 3 for a state-of-the-art agent.This demonstrates that the perceived variance is not necessarily inherent to RL.Instead, it may be addressed via simple modifications and we argue that developing low-variance agents is an important goal for the RL community. | IS HIGH VARIANCE UNAVOIDABLE IN RL? A CASE STUDY IN CONTINUOUS CONTROL |
d263831010 | Recently, pre-trained foundation models have enabled significant advancements in multiple fields.In molecular machine learning, however, where datasets are often hand-curated, and hence typically small, the lack of datasets with labeled features, and codebases to manage those datasets, has hindered the development of foundation models.In this work, we present seven novel datasets categorized by size into three distinct categories: ToyMix, LargeMix and UltraLarge.These datasets push the boundaries in both the scale and the diversity of supervised labels for molecular learning.They cover nearly 100 million molecules and over 3000 sparsely defined tasks, totaling more than 13 billion individual labels of both quantum and biological nature.In comparison, our datasets contain 300 times more data points than the widely used OGB-LSC PCQM4Mv2 dataset, and 13 times more than the quantum-only QM1B dataset.In addition, to support the development of foundational models based on our proposed datasets, we present the Graphium graph machine learning library which simplifies the process of building and training molecular machine learning models for multi-task and multi-level molecular datasets.Finally, we present a range of baseline results as a starting point of multi-task and multi-level training on these datasets.Empirically, we observe that performance on low-resource biological datasets show improvement by also training on large amounts of quantum data.This indicates that there may be potential in multi-task and multi-level training of a foundation model and fine-tuning it to resource-constrained downstream tasks. | TOWARDS FOUNDATIONAL MODELS FOR MOLECULAR LEARNING ON LARGE-SCALE MULTI-TASK DATASETS |
d3609219 | We present Adaptive Memory Networks (AMN) that processes input-question pairs to dynamically construct a network architecture optimized for lower inference times for Question Answering (QA) tasks. AMN processes the input story to extract entities and stores them in memory banks. Starting from a single bank, as the number of input entities increases, AMN learns to create new banks as the entropy in a single bank becomes too high. Hence, after processing an input-question(s) pair, the resulting network represents a hierarchical structure where entities are stored in different banks, distanced by question relevance. At inference, one or few banks are used, creating a tradeoff between accuracy and performance.AMN is enabled by dynamic networks that allow input dependent network creation and efficiency in dynamic mini-batching as well as our novel bank controller that allows learning discrete decision making with high accuracy. In our results, we demonstrate that AMN learns to create variable depth networks depending on task complexity and reduces inference times for QA tasks. | Adaptive Memory Networks |
d238407831 | Climate change is a major threat to humanity, and the actions required to prevent its catastrophic consequences include changes in both policy-making and individual behaviour. However, taking action requires understanding the effects of climate change, even though they may seem abstract and distant. Projecting the potential consequences of extreme climate events such as flooding in familiar places can help make the abstract impacts of climate change more concrete and encourage action. As part of a larger initiative to build a website that projects extreme climate events onto user-chosen photos, we present our solution to simulate photo-realistic floods on authentic images. To address this complex task in the absence of suitable training data, we propose ClimateGAN, a model that leverages both simulated and real data for unsupervised domain adaptation and conditional image generation. In this paper, we describe the details of our framework, thoroughly evaluate components of our architecture and demonstrate that our model is capable of robustly generating photo-realistic flooding. * Corresponding authors -[schmidtv, luccionis]@mila.quebec Preprint. Under review. | ClimateGAN: Raising Climate Change Awareness by Generating Images of Floods |
d259095535 | AlphaZero-type algorithms may stop improving on single-player tasks in case the value network guiding the tree search is unable to approximate the outcome of an episode sufficiently well. One technique to address this problem is transforming the single-player task through self-competition. The main idea is to compute a scalar baseline from the agent's historical performances and to reshape an episode's reward into a binary output, indicating whether the baseline has been exceeded or not. However, this baseline only carries limited information for the agent about strategies how to improve. We leverage the idea of self-competition and directly incorporate a historical policy into the planning process instead of its scalar performance. Based on the recently introduced Gumbel AlphaZero (GAZ), we propose our algorithm GAZ 'Play-to-Plan' (GAZ PTP), in which the agent learns to find strong trajectories by planning against possible strategies of its past self. We show the effectiveness of our approach in two well-known combinatorial optimization problems, the Traveling Salesman Problem and the Job-Shop Scheduling Problem. With only half of the simulation budget for search, GAZ PTP consistently outperforms all selected single-player variants of GAZ. | POLICY-BASED SELF-COMPETITION FOR PLANNING PROBLEMS |
d253224188 | One-shot coreset selection aims to select a representative subset of the training data, given a pruning rate, that can later be used to train future models while retaining high accuracy. State-of-the-art coreset selection methods pick the highest importance examples based on an importance metric and are found to perform well at low pruning rates. However, at high pruning rates, they suffer from a catastrophic accuracy drop, performing worse than even random sampling. This paper explores the reasons behind this accuracy drop both theoretically and empirically. We first propose a novel metric to measure the coverage of a dataset on a specific distribution by extending the classical geometric set cover problem to a distribution cover problem. This metric helps explain why coresets selected by SOTA methods at high pruning rates perform poorly compared to random sampling because of worse data coverage. We then propose a novel one-shot coreset selection method, Coverage-centric Coreset Selection (CCS), that jointly considers overall data coverage upon a distribution as well as the importance of each example. We evaluate CCS on five datasets and show that, at high pruning rates (e.g., 90%), it achieves significantly better accuracy than previous SOTA methods (e.g., at least 19.56% higher on CIFAR10) as well as random selection (e.g., 7.04% higher on CIFAR10) and comparable accuracy at low pruning rates. We make our code publicly available at GitHub 1 . | COVERAGE-CENTRIC CORESET SELECTION FOR HIGH PRUNING RATES |
d258298441 | Feature distillation makes the student mimic the intermediate features of the teacher. Nearly all existing feature-distillation methods use L2 distance or its slight variants as the distance metric between teacher and student features. However, while L2 distance is isotropic w.r.t. all dimensions, the neural network's operation on different dimensions is usually anisotropic, i.e., perturbations with the same 2-norm but in different dimensions of intermediate features lead to changes in the final output with largely different magnitude. Considering this, we argue that the similarity between teacher and student features should not be measured merely based on their appearance (i.e., L2 distance), but should, more importantly, be measured by their difference in function, namely how later layers of the network will read, decode, and process them. Therefore, we propose Function-Consistent Feature Distillation (FCFD), which explicitly optimizes the functional similarity between teacher and student features. The core idea of FCFD is to make teacher and student features not only numerically similar, but more importantly produce similar outputs when fed to the later part of the same network. With FCFD, the student mimics the teacher more faithfully and learns more from the teacher. Extensive experiments on image classification and object detection demonstrate the superiority of FCFD to existing methods. Furthermore, we can combine FCFD with many existing methods to obtain even higher accuracy. Our codes are available at https://github.com/LiuDongyang6/FCFD. | FUNCTION-CONSISTENT FEATURE DISTILLATION |
d231924830 | Spatial memory, or the ability to remember and recall specific locations and objects, is central to autonomous agents' ability to carry out tasks in real environments. However, most existing artificial memory modules are not very adept at storing spatial information. We propose a parameter-free module, Egospheric Spatial Memory (ESM), which encodes the memory in an ego-sphere around the agent, enabling expressive 3D representations. ESM can be trained end-to-end via either imitation or reinforcement learning, and improves both training efficiency and final performance against other memory baselines on both drone and manipulator visuomotor control tasks. The explicit egocentric geometry also enables us to seamlessly combine the learned controller with other non-learned modalities, such as local obstacle avoidance. We further show applications to semantic segmentation on the ScanNet dataset, where ESM naturally combines image-level and map-level inference modalities. Through our broad set of experiments, we show that ESM provides a general computation graph for embodied spatial reasoning, and the module forms a bridge between real-time mapping systems and differentiable memory architectures. Implementation at: https://github.com/ivy-dl/memory. -to-real via sim-to-sim: Dataefficient robotic grasping via randomized-to-canonical adaptation networks. In . Neural rgb (r) d sensing: Depth and uncertainty from a video camera. In , et al. Learning to navigate in cities without a map. , et al. Human-level control through deep reinforcement learning. nature, 518(7540):529-533, 2015. The sensory ego-sphere as a short-term memory for humanoids. In | END-TO-END EGOSPHERIC SPATIAL MEMORY |
d252595765 | Feature selection is the problem of selecting a subset of features for a machine learning model that maximizes model quality subject to a budget constraint. For neural networks, prior methods, including those based on ℓ 1 regularization, attention, and other techniques, typically select the entire feature subset in one evaluation round, ignoring the residual value of features during selection, i.e., the marginal contribution of a feature given that other features have already been selected. We propose a feature selection algorithm called Sequential Attention that achieves state-of-the-art empirical results for neural networks. This algorithm is based on an efficient one-pass implementation of greedy forward selection and uses attention weights at each step as a proxy for feature importance. We give theoretical insights into our algorithm for linear regression by showing that an adaptation to this setting is equivalent to the classical Orthogonal Matching Pursuit (OMP) algorithm, and thus inherits all of its provable guarantees. Our theoretical and empirical analyses offer new explanations towards the effectiveness of attention and its connections to overparameterization, which may be of independent interest. | SEQUENTIAL ATTENTION FOR FEATURE SELECTION |
d58006571 | The adversarial training procedure proposed byMadry et al. (2018)is one of the most effective methods to defend against adversarial examples in deep neural networks (DNNs). In our paper, we shed some lights on the practicality and the hardness of adversarial training by showing that the effectiveness (robustness on test set) of adversarial training has a strong correlation with the distance between a test point and the manifold of training data embedded by the network. Test examples that are relatively far away from this manifold are more likely to be vulnerable to adversarial attacks. Consequentially, an adversarial training based defense is susceptible to a new class of attacks, the "blind-spot attack", where the input images reside in "blind-spots" (low density regions) of the empirical distribution of training data but is still on the ground-truth data manifold. For MNIST, we found that these blind-spots can be easily found by simply scaling and shifting image pixel values. Most importantly, for large datasets with high dimensional and complex data manifold (CIFAR, ImageNet, etc), the existence of blind-spots in adversarial training makes defending on any valid test examples difficult due to the curse of dimensionality and the scarcity of training data. Additionally, we find that blind-spots also exist on provable defenses including and(Sinha et al., 2018)because these trainable robustness certificates can only be practically optimized on a limited set of training data. | THE LIMITATIONS OF ADVERSARIAL TRAINING AND THE BLIND-SPOT ATTACK |
d3463636 | Recent studies show that widely used deep neural networks (DNNs) are vulnerable to carefully crafted adversarial examples. Many advanced algorithms have been proposed to generate adversarial examples by leveraging the L p distance for penalizing perturbations. Researchers have explored different defense methods to defend against such adversarial attacks. While the effectiveness of L p distance as a metric of perceptual quality remains an active research area, in this paper we will instead focus on a different type of perturbation, namely spatial transformation, as opposed to manipulating the pixel values directly as in prior works. Perturbations generated through spatial transformation could result in large L p distance measures, but our extensive experiments show that such spatially transformed adversarial examples are perceptually realistic and more difficult to defend against with existing defense systems. This potentially provides a new direction in adversarial example generation and the design of corresponding defenses. We visualize the spatial transformation based perturbation for different examples and show that our technique can produce realistic adversarial examples with smooth image deformation. Finally, we visualize the attention of deep networks with different types of adversarial examples to better understand how these examples are interpreted. * indicates equal contributions | SPATIALLY TRANSFORMED ADVERSARIAL EXAMPLES |
d254198961 | Poor sample efficiency continues to be the primary challenge for deployment of deep Reinforcement Learning (RL) algorithms for real-world applications, and in particular for visuo-motor control. Model-based RL has the potential to be highly sample efficient by concurrently learning a world model and using synthetic rollouts for planning and policy improvement. However, in practice, sample-efficient learning with model-based RL is bottlenecked by the exploration challenge. In this work, we find that leveraging just a handful of demonstrations can dramatically improve the sample-efficiency of model-based RL. Simply appending demonstrations to the interaction dataset, however, does not suffice. We identify key ingredients for leveraging demonstrations in model learning -policy pretraining, targeted exploration, and oversampling of demonstration data -which forms the three phases of our model-based RL framework. We empirically study three complex visuo-motor control domains and find that our method is 160% − 250% more successful in completing sparse reward tasks compared to prior approaches in the low data regime (100K interaction steps, 5 demonstrations). Code and videos are available at: https://nicklashansen.github.io/modemrl. arXiv:2212.05698v1 [cs.LG] 12 Dec 2022 Preprint Figure 2. Our framework (MoDem) consists of three phases:(1) a policy pretraining phase where representation and policy is trained on a handful of demonstrations via BC, (2) a seeding phase where the pretrained policy is used to generate rollouts for targeted model learning, and (3) an interactive learning phase where the model iteratively collects new rollouts and is trained with data from all three phases. Crucially, we aggressively oversample demonstration data for model learning, regularize the model using data augmentation, and reuse weights across phases. sg: stop-gradient operator.Shah & Kumar, 2021;Zhan et al., 2020). In this work, we propose a new framework to accelerate model-based RL algorithms with demonstrations. On a suite of challenging visuo-motor control tasks, we find that our method can train policies that are approx. 160% − 250% more successful than prior state-of-the-art (SOTA) baselines. visuomotor policies. . R3m: A universal visual representation for robot manipulation. ArXiv, abs/2203.12601, 2022. | MODEM: ACCELERATING VISUAL MODEL-BASED REINFORCEMENT LEARNING WITH DEMONSTRATIONS |
d249209856 | SHAP explanations aim at identifying which features contribute the most to the difference in model prediction at a specific input versus a background distribution. Recent studies have shown that they can be manipulated by malicious adversaries to produce arbitrary desired explanations. However, existing attacks focus solely on altering the black-box model itself. In this paper, we propose a complementary family of attacks that leave the model intact and manipulate SHAP explanations using stealthily biased sampling of the data points used to approximate expectations w.r.t the background distribution. In the context of fairness audit, we show that our attack can reduce the importance of a sensitive feature when explaining the difference in outcomes between groups, while remaining undetected. These results highlight the manipulability of SHAP explanations and encourage auditors to treat post-hoc explanations with skepticism. Recently, several studies reported that such a manipulation is possible, e.g. by modifying the Preprint. Under review. | Fooling SHAP with Stealthily Biased Sampling |
d67855499 | While Bayesian neural networks (BNNs) have drawn increasing attention, their posterior inference remains challenging, due to the high-dimensional and overparameterized nature. Recently, several highly flexible and scalable variational inference procedures based on the idea of particle optimization have been proposed. These methods directly optimize a set of particles to approximate the target posterior. However, their application to BNNs often yields sub-optimal performance, as they have a particular failure mode on over-parameterized models. In this paper, we propose to solve this issue by performing particle optimization directly in the space of regression functions. We demonstrate through extensive experiments that our method successfully overcomes this issue, and outperforms strong baselines in a variety of tasks including prediction, defense against adversarial examples, and reinforcement learning. | FUNCTION SPACE PARTICLE OPTIMIZATION FOR BAYESIAN NEURAL NETWORKS |
d258967589 | Preference-based Reinforcement Learning (PbRL) is a paradigm in which an RL agent learns to optimize a task using pair-wise preference-based feedback over trajectories, rather than explicit reward signals. While PbRL has demonstrated practical success in fine-tuning language models, existing theoretical work focuses on regret minimization and fails to capture most of the practical frameworks. In this study, we fill in such a gap between theoretical PbRL and practical algorithms by proposing a theoretical reward-agnostic PbRL framework where exploratory trajectories that enable accurate learning of hidden reward functions are acquired before collecting any human feedback. Theoretical analysis demonstrates that our algorithm requires less human feedback for learning the optimal policy under preference-based models with linear parameterization and unknown transitions, compared to the existing theoretical literature. Specifically, our framework can incorporate linear and low-rank MDPs with efficient sample complexity. Additionally, we investigate reward-agnostic RL with action-based comparison feedback and introduce an efficient querying algorithm tailored to this scenario. | Provable Reward-Agnostic Preference-Based Reinforcement Learning |
d257219472 | With adversarial or otherwise normal prompts, existing large language models (LLM) can be pushed to generate toxic discourses. One way to reduce the risk of LLMs generating undesired discourses is to alter the training of the LLM. This can be very restrictive due to demanding computation requirements. Other methods rely on rule-based or prompt-based token elimination, which are limited as they dismiss future tokens and the overall meaning of the complete discourse. Here, we center detoxification on the probability that the finished discourse is ultimately considered toxic. That is, at each point, we advise against token selections proportional to how likely a finished text from this point will be toxic. To this end, we formally extend the dead-end theory from the recent reinforcement learning (RL) literature to also cover uncertain outcomes. Our approach, called rectification, utilizes a separate but significantly smaller model for detoxification, which can be applied to diverse LLMs as long as they share the same vocabulary. Importantly, our method does not require access to the internal representations of the LLM, but only the token probability distribution at each decoding step. This is crucial as many LLMs today are hosted in servers and only accessible through APIs. When applied to various LLMs, including GPT-3, our approach significantly improves the generated discourse compared to the base LLMs and other techniques in terms of both the overall language and detoxification performance. | SYSTEMATIC RECTIFICATION OF LANGUAGE MODELS VIA DEAD-END ANALYSIS |
d258180544 | Contemporary applications of machine learning in two-team e-sports and the superior expressivity of multi-agent generative adversarial networks raise important and overlooked theoretical questions regarding optimization in two-team games. Formally, two-team zero-sum games are defined as multi-player games where players are split into two competing sets of agents, each experiencing a utility identical to that of their teammates and opposite to that of the opposing team. We focus on the solution concept of Nash equilibria (NE). We first show that computing NE for this class of games is hard for the complexity class CLS. To further examine the capabilities of online learning algorithms in games with full-information feedback, we propose a benchmark of a simple -yet nontrivial-family of such games. These games do not enjoy the properties used to prove convergence for relevant algorithms. In particular, we use a dynamical systems perspective to demonstrate that gradient descent-ascent, its optimistic variant, optimistic multiplicative weights update, and extra gradient fail to converge (even locally) to a Nash equilibrium. On a brighter note, we propose a first-order method that leverages control theory techniques and under some conditions enjoys last-iterate local convergence to a Nash equilibrium. We also believe our proposed method is of independent interest for general min-max optimization. | Towards convergence to Nash equilibria in two-team zero-sum games |
d210473083 | Most generative models of audio directly generate samples in one of two domains: time or frequency. While sufficient to express any signal, these representations are inefficient, as they do not utilize existing knowledge of how sound is generated and perceived. A third approach (vocoders/synthesizers) successfully incorporates strong domain knowledge of signal processing and perception, but has been less actively researched due to limited expressivity and difficulty integrating with modern auto-differentiation-based machine learning methods. In this paper, we introduce the Differentiable Digital Signal Processing (DDSP) library, which enables direct integration of classic signal processing elements with deep learning methods. Focusing on audio synthesis, we achieve high-fidelity generation without the need for large autoregressive models or adversarial losses, demonstrating that DDSP enables utilizing strong inductive biases without losing the expressive power of neural networks. Further, we show that combining interpretable modules permits manipulation of each separate model component, with applications such as independent control of pitch and loudness, realistic extrapolation to pitches not seen during training, blind dereverberation of room acoustics, transfer of extracted room acoustics to new environments, and transformation of timbre between disparate sources. In short, DDSP enables an interpretable and modular approach to generative modeling, without sacrificing the benefits of deep learning. The library is publicly available 1 and we welcome further contributions from the community and domain experts. | DDSP: DIFFERENTIABLE DIGITAL SIGNAL PROCESSING |
d51559 | Recurrent neural networks are a powerful tool for modeling sequential data, but the dependence of each timestep's computation on the previous timestep's output limits parallelism and makes RNNs unwieldy for very long sequences. We introduce quasi-recurrent neural networks (QRNNs), an approach to neural sequence modeling that alternates convolutional layers, which apply in parallel across timesteps, and a minimalist recurrent pooling function that applies in parallel across channels. Despite lacking trainable recurrent layers, stacked QRNNs have better predictive accuracy than stacked LSTMs of the same hidden size. Due to their increased parallelism, they are up to 16 times faster at train and test time. Experiments on language modeling, sentiment classification, and character-level neural machine translation demonstrate these advantages and underline the viability of QRNNs as a basic building block for a variety of sequence tasks. * Equal contribution | QUASI-RECURRENT NEURAL NETWORKS |
d252734944 | Equivariance guarantees that a model's predictions capture key symmetries in data. When an image is translated or rotated, an equivariant model's representation of that image will translate or rotate accordingly. The success of convolutional neural networks has historically been tied to translation equivariance directly encoded in their architecture. The rising success of vision transformers, which have no explicit architectural bias towards equivariance, challenges this narrative and suggests that augmentations and training data might also play a significant role in their performance. In order to better understand the role of equivariance in recent vision models, we introduce the Lie derivative, a method for measuring equivariance with strong mathematical foundations and minimal hyperparameters. Using the Lie derivative, we study the equivariance properties of hundreds of pretrained models, spanning CNNs, transformers, and Mixer architectures. The scale of our analysis allows us to separate the impact of architecture from other factors like model size or training method. Surprisingly, we find that many violations of equivariance can be linked to spatial aliasing in ubiquitous network layers, such as pointwise non-linearities, and that as models get larger and more accurate they tend to display more equivariance, regardless of architecture. For example, transformers can be more equivariant than convolutional neural networks after training. * Equal contribution. | The Lie Derivative for Measuring Learned Equivariance |
d245668925 | Text adventure games present unique challenges to reinforcement learning methods due to their combinatorially large action spaces and sparse rewards. The interplay of these two factors is particularly demanding because large action spaces require extensive exploration, while sparse rewards provide limited feedback. This work proposes to tackle the explore-vs-exploit dilemma using a multi-stage approach that explicitly disentangles these two strategies within each episode. Our algorithm, called eXploit-Then-eXplore (XTX), begins each episode using an exploitation policy that imitates a set of promising trajectories from the past, and then switches over to an exploration policy aimed at discovering novel actions that lead to unseen state spaces. This policy decomposition allows us to combine global decisions about which parts of the game space to return to with curiosity-based local exploration in that space, motivated by how a human may approach these games. Our method significantly outperforms prior approaches by 27% and 11% average normalized score over 12 games from the Jericho benchmark in both deterministic and stochastic settings, respectively. On the game of Zork1, in particular, XTX obtains a score of 103, more than a 2x improvement over prior methods, and pushes past several known bottlenecks in the game that have plagued previous state-of-the-art methods. 1 | MULTI-STAGE EPISODIC CONTROL FOR STRATEGIC EXPLORATION IN TEXT GAMES |
d222378250 | Self-supervised learning has emerged as a strategy to reduce the reliance on costly supervised signal by pretraining representations only using unlabeled data. These methods combine heuristic proxy classification tasks with data augmentations and have achieved significant success, but our theoretical understanding of this success remains limited. In this paper we analyze self-supervised representation learning using a causal framework. We show how data augmentations can be more effectively utilized through explicit invariance constraints on the proxy classifiers employed during pretraining. Based on this, we propose a novel selfsupervised objective, Representation Learning via Invariant Causal Mechanisms (RELIC), that enforces invariant prediction of proxy targets across augmentations through an invariance regularizer which yields improved generalization guarantees. Further, using causality we generalize contrastive learning, a particular kind of self-supervised method, and provide an alternative theoretical explanation for the success of these methods. Empirically, RELIC significantly outperforms competing methods in terms of robustness and out-of-distribution generalization on ImageNet, while also significantly outperforming these methods on Atari achieving above human-level performance on 51 out of 57 games. | REPRESENTATION LEARNING VIA INVARIANT CAUSAL MECHANISMS |
d108308108 | A key challenge in reinforcement learning (RL) is environment generalization: a policy trained to solve a task in one environment often fails to solve the same task in a slightly different test environment. A common approach to improve interenvironment transfer is to learn policies that are invariant to the distribution of testing environments. However, we argue that instead of being invariant, the policy should identify the specific nuances of an environment and exploit them to achieve better performance. In this work, we propose the "Environment-Probing" Interaction (EPI) policy, a policy that probes a new environment to extract an implicit understanding of that environment's behavior. Once this environmentspecific information is obtained, it is used as an additional input to a task-specific policy that can now perform environment-conditioned actions to solve a task. To learn these EPI-policies, we present a reward function based on transition predictability. Specifically, a higher reward is given if the trajectory generated by the EPI-policy can be used to better predict transitions. We experimentally show that EPI-conditioned task-specific policies significantly outperform commonly used policy generalization methods on novel testing environments. | ENVIRONMENT PROBING INTERACTION POLICIES |
d258564352 | The exploration problem is one of the main challenges in deep reinforcement learning (RL). Recent promising works tried to handle the problem with populationbased methods, which collect samples with diverse behaviors derived from a population of different exploratory policies. Adaptive policy selection has been adopted for behavior control. However, the behavior selection space is largely limited by the predefined policy population, which further limits behavior diversity. In this paper, we propose a general framework called Learnable Behavioral Control (LBC) to address the limitation, which a) enables a significantly enlarged behavior selection space via formulating a hybrid behavior mapping from all policies; b) constructs a unified learnable process for behavior selection. We introduce LBC into distributed off-policy actor-critic methods and achieve behavior control via optimizing the selection of the behavior mappings with bandit-based metacontrollers. Our agents have achieved 10077.52% mean human normalized score and surpassed 24 human world records within 1B training frames in the Arcade Learning Environment, which demonstrates our significant state-of-the-art (SOTA) performance without degrading the sample efficiency. | LEARNABLE BEHAVIOR CONTROL: BREAKING ATARI HUMAN WORLD RECORDS VIA SAMPLE-EFFICIENT BEHAVIOR SELECTION |
d251066616 | The problem of end-to-end learning of a communication system using an autoencoder -consisting of an encoder, channel, and decoder modeled using neural networks -has recently been shown to be an effective approach. A challenge faced in the practical adoption of this learning approach is that under changing channel conditions (e.g. a wireless link), it requires frequent retraining of the autoencoder in order to maintain a low decoding error rate. Since retraining is both time consuming and requires a large number of samples, it becomes impractical when the channel distribution is changing quickly. We propose to address this problem using a fast and sample-efficient (few-shot) domain adaptation method that does not change the encoder and decoder networks. Different from conventional training-time unsupervised or semi-supervised domain adaptation, here we have a trained autoencoder from a source distribution that we want to adapt (at test time) to a target distribution using only a small labeled dataset, and no unlabeled data. We focus on a generative channel model based on the Gaussian mixture density network (MDN), and propose a regularized, parameter-efficient adaptation of the MDN using a set of affine transformations. The learned affine transformations are then used to design an optimal transformation at the decoder input to compensate for the distribution shift, and effectively present to the decoder inputs close to the source distribution. Experiments on many simulated distribution changes common to the wireless setting, and a real mmWave FPGA testbed demonstrate the effectiveness of our method at adaptation using very few target domain samples 1 . | FEW-SHOT DOMAIN ADAPTATION FOR END-TO-END COMMUNICATION |
d252993068 | While Mixed-integer linear programming (MILP) is NP-hard in general, practical MILP has received roughly 100-fold speedup in the past twenty years. Still, many classes of MILPs quickly become unsolvable as their sizes increase, motivating researchers to seek new acceleration techniques for MILPs. With deep learning, they have obtained strong empirical results, and many results were obtained by applying graph neural networks (GNNs) to making decisions in various stages of MILP solution processes. This work discovers a fundamental limitation: there exist feasible and infeasible MILPs that all GNNs will, however, treat equally, indicating GNN's lacking power to express general MILPs. Then, we show that, by restricting the MILPs to unfoldable ones or by adding random features, there exist GNNs that can reliably predict MILP feasibility, optimal objective values, and optimal solutions up to prescribed precision. We conducted small-scale numerical experiments to validate our theoretical findings. | ON REPRESENTING MIXED-INTEGER LINEAR PROGRAMS BY GRAPH NEURAL NETWORKS |
d238419564 | For many differentially private algorithms, such as the prominent noisy stochastic gradient descent (DP-SGD), the analysis needed to bound the privacy leakage of a single training run is well understood. However, few studies have reasoned about the privacy leakage resulting from the multiple training runs needed to fine tune the value of the training algorithm's hyperparameters. In this work, we first illustrate how simply setting hyperparameters based on non-private training runs can leak private information. Motivated by this observation, we then provide privacy guarantees for hyperparameter search procedures within the framework of Renyi Differential Privacy. Our results improve and extend the work of Liu and Talwar (STOC 2019). Our analysis supports our previous observation that tuning hyperparameters does indeed leak private information, but we prove that, under certain assumptions, this leakage is modest, as long as each candidate training run needed to select hyperparameters is itself differentially private. | HYPERPARAMETER TUNING WITH RENYI DIFFERENTIAL PRIVACY |
d235613557 | Understanding human behavior from observed data is critical for transparency and accountability in decision-making.Consider real-world settings such as healthcare, in which modeling a decision-maker's policy is challenging-with no access to underlying states, no knowledge of environment dynamics, and no allowance for live experimentation.We desire learning a data-driven representation of decisionmaking behavior that (1) inheres transparency by design, (2) accommodates partial observability, and (3) operates completely offline.To satisfy these key criteria, we propose a novel model-based Bayesian method for interpretable policy learning ("INTERPOLE") that jointly estimates an agent's (possibly biased) belief-update process together with their (possibly suboptimal) belief-action mapping.Through experiments on both simulated and real-world data for the problem of Alzheimer's disease diagnosis, we illustrate the potential of our approach as an investigative device for auditing, quantifying, and understanding human decision-making behavior. | EXPLAINING BY IMITATING: UNDERSTANDING DECISIONS BY INTERPRETABLE POLICY LEARNING |
d258947197 | This paper investigates a missing feature imputation problem for graph learning tasks. Several methods have previously addressed learning tasks on graphs with missing features. However, in cases of high rates of missing features, they were unable to avoid significant performance degradation. To overcome this limitation, we introduce a novel concept of channel-wise confidence in a node feature, which is assigned to each imputed channel feature of a node for reflecting certainty of the imputation. We then design pseudo-confidence using the channel-wise shortest path distance between a missing-feature node and its nearest known-feature node to replace unavailable true confidence in an actual learning process. Based on the pseudo-confidence, we propose a novel feature imputation scheme that performs channel-wise inter-node diffusion and node-wise inter-channel propagation. The scheme can endure even at an exceedingly high missing rate (e.g., 99.5%) and it achieves state-of-the-art accuracy for both semi-supervised node classification and link prediction on various datasets containing a high rate of missing features. Codes are available at https://github.com/daehoum1/pcfi. Michael Bronstein. On the unreasonable effectiveness of feature propagation in learning on graphs with missing node features. arXiv preprint arXiv: | CONFIDENCE-BASED FEATURE IMPUTATION FOR GRAPHS WITH PARTIALLY KNOWN FEATURES |
d239050418 | Robustness to distribution shifts is critical for deploying machine learning models in the real world. Despite this necessity, there has been little work in defining the underlying mechanisms that cause these shifts and evaluating the robustness of algorithms across multiple, different distribution shifts. To this end, we introduce a framework that enables fine-grained analysis of various distribution shifts. We provide a holistic analysis of current state-of-the-art methods by evaluating 19 distinct methods grouped into five categories across both synthetic and real-world datasets. Overall, we train more than 85K models. Our experimental framework can be easily extended to include new methods, shifts, and datasets. We find, unlike previous work (Gulrajani & Lopez-Paz, 2021), that progress has been made over a standard ERM baseline; in particular, pretraining and augmentations (learned or heuristic) offer large gains in many cases. However, the best methods are not consistent over different datasets and shifts. . Model-agnostic meta-learning for fast adaptation of deep networks. In . Imagenet-trained cnns are biased towards texture; increasing shape bias improves accuracy and robustness. In Achieving robustness in the wild via adversarial mixing with disentangled representations. In | A FINE-GRAINED ANALYSIS ON DISTRIBUTION SHIFT |
d263829511 | Figure 1: Comparison between our method, Linear Autoregressive Similarity Index (LASI), and Learned Perceptual Image Patch Similarity (LPIPS) from Zhang et al. (2018).Our method solves a weighted least squares (WLS) problem to compute perceptual embeddings, at inference time, with no prior training or neural networks.LPIPS uses pre-trained features from deep models as image embeddings and trains on human annotated data. | THE UNREASONABLE EFFECTIVENESS OF LINEAR PREDICTION AS A PERCEPTUAL METRIC |
d211082795 | Local robustness ensures that a model classifies all inputs within an -ball consistently, which precludes various forms of adversarial inputs. In this paper, we present a fast procedure for checking local robustness in feed-forward neural networks with piecewise linear activation functions. The key insight is that such networks partition the input space into a polyhedral complex such that the network is linear inside each polyhedral region; hence, a systematic search for decision boundaries within the regions around a given input is sufficient for assessing robustness. Crucially, we show how these regions can be analyzed using geometric projections instead of expensive constraint solving, thus admitting an efficient, highly-parallel GPU implementation at the price of incompleteness, which can be addressed by falling back on prior approaches. Empirically, we find that incompleteness is not often an issue, and that our method performs one to two orders of magnitude faster than existing robustness-certification techniques based on constraint solving. | Fast Geometric Projections for Local Robustness Certification |
d233306870 | Machine learning systems typically assume that the distributions of training and test sets match closely. However, a critical requirement of such systems in the real world is their ability to generalize to unseen domains. Here, we propose an inter-domain gradient matching objective that targets domain generalization by maximizing the inner product between gradients from different domains. Since direct optimization of the gradient inner product can be computationally prohibitive -it requires computation of second-order derivatives --we derive a simpler first-order algorithm named Fish that approximates its optimization. We perform experiments on the WILDS benchmark, which captures distribution shift in the real world, as well as the DOMAINBED benchmark that focuses more on syntheticto-real transfer. Our method produces competitive results on both benchmarks, demonstrating its effectiveness across a wide range of domain generalization tasks. | Gradient Matching for Domain Generalization |
d252683303 | We study the task of prompting large-scale language models to perform multistep reasoning. Existing work shows that when prompted with a chain of thoughts (CoT), sequences of short sentences describing intermediate reasoning steps towards a final answer, large language models can generate new reasoning chains and predict answers for new inputs. A central question is which reasoning examples make the most effective prompts. In this work, we propose complexitybased prompting, a simple and effective example selection scheme for multi-step reasoning. We show that prompts with higher reasoning complexity, i.e., chains with more reasoning steps, achieve substantially better performance on multistep reasoning tasks over strong baselines. We further extend our complexitybased criteria from prompting (selecting inputs) to decoding (selecting outputs), where we sample multiple reasoning chains from the model, then choose the majority of generated answers from complex reasoning chains (over simple chains). When used to prompt GPT-3 and Codex, our approach substantially improves multi-step reasoning accuracy and achieves new state-of-the-art (SOTA) performance on three math benchmarks (GSM8K, MultiArith, and MathQA) and two BigBenchHard tasks (Date Understanding and Penguins), with an average +5.3 and up to +18 accuracy improvements. Compared with existing example selection schemes like manual tuning or retrieval-based selection, selection based on reasoning complexity is intuitive, easy to implement, and annotation-efficient. Further results demonstrate the robustness of performance gains from complex prompts under format perturbation and distribution shift. | COMPLEXITY-BASED PROMPTING FOR MULTI-STEP REASONING |
d2483823 | Deep Convolutional Neuronal Networks (DCNNs) are showing remarkable performance on many computer vision tasks. Due to their large parameter space, they require many labeled samples when trained in a supervised setting. The costs of annotating data manually can render the use of DCNNs infeasible. We present a novel framework called RenderGAN that can generate large amounts of realistic, labeled images by combining a 3D model and the Generative Adversarial Network framework. In our approach, image augmentations (e.g., lighting, background, and detail) are learned from unlabeled data such that the generated images are strikingly realistic while preserving the labels known from the 3D model. We apply the RenderGAN framework to generate images of barcode-like markers that are attached to honeybees. Training a DCNN on data generated by the RenderGAN yields considerably better performance than training it on various baselines. | RenderGAN: Generating Realistic Labeled Data |
d248965297 | Human language is grounded on multimodal knowledge including visual knowledge like colors, sizes, and shapes. However, current large-scale pre-trained language models rely on text-only self-supervised training with massive text data, which precludes them from utilizing relevant visual information when necessary. To address this, we propose a novel pre-training framework, named VALM, to Visually-augment text tokens with retrieved relevant images for Language Modeling. Specifically, VALM builds on a novel latent text-image alignment method via an image retrieval module to fetch corresponding images given a textual context. With the visually-augmented context, VALM uses a visual knowledge fusion layer to enable multimodal grounded language modeling by attending to both text context and visual knowledge in images. We evaluate VALM on various visual knowledge-intensive commonsense reasoning tasks, which require visual information to excel. The experimental results illustrate that VALM outperforms all strong language-only and vision-language baselines with substantial gains in reasoning object commonsense including color, size, and shape. Our code is available at https://github.com/Victorwz/VaLM. . Efficient large-scale language model training on gpu clusters using megatron-lm. . fairseq: A fast, extensible toolkit for sequence modeling. In NAACL, 2019. . Erica: Improving entity and relation understanding for pre-trained language models via contrastive learning. In ACL, 2021. | VISUALLY-AUGMENTED LANGUAGE MODELING |
d256868350 | Today, ground-truth generation uses data sets annotated by cloud-based annotation services. These services rely on human annotation, which can be prohibitively expensive. In this paper, we consider the problem of hybrid human-machine labeling, which trains a classifier to accurately auto-label part of the data set. However, training the classifier can be expensive too. We propose an iterative approach that minimizes total overall cost by, at each step, jointly determining which samples to label using humans and which to label using the trained classifier. We validate our approach on well known public data sets such as Fashion-MNIST, CIFAR-10, CIFAR-100, and ImageNet. In some cases, our approach has 6× lower overall cost relative to human labeling the entire data set, and is always cheaper than the cheapest competing strategy.Hybrid Human-machine Annotations. In this paper, we explore using a hybrid human-machine approach to reduce annotation costs (in $) where humans only annotate a subset of the data items; a machine learning model trained on this annotated data annotates the rest. The accuracy of a model trained on a subset of the data set will typically be lower than that of human annotators. However, a user of an annotation service might choose to avail of this trade-off if (a) targeting a slightly lower annotation quality can significantly reduce costs, or (b) the cost of training a model to a higher accuracy is itself prohibitive. Consequently, this paper focuses on the design of a hybrid human-machine annotation scheme that minimizes the overall cost of annotating the entire data set (including the cost of training the model) while ensuring that the overall annotation accuracy, relative to human annotations, is higher than a pre-specified target (e.g., 95%).Challenges. In this paper, we consider a specific annotation task, multi-class labeling. We assume that the user of an annotation service provides a set X of data to be labeled and a classifier D to use for machine labeling. Then, the goal is to find a subset B ⊂ X human-labeled samples to train D, and use the classifier to label the rest, minimizing total cost while ensuring the target accuracy. A straw man approach might seek to predict human-labeled subset B in a single shot. This is hard to do because it depends on several factors: (a) the classifier architecture and how much accuracy it can achieve, (b) how "hard" the dataset is, (c) the cost of training and labeling, and (d) the target accuracy. Complex models may provide a high accuracy, their training costs may be too high and potentially offset the gains obtained through machine-generated annotations. Some data-points in a dataset are more informative as compared to the rest from a model training perspective. Identifying the "right" data subset for human-vs. machine-labeling can minimize the total labeling cost.Approach. In this paper we propose a novel technique, MCAL 1 Sample Data ML Label FIGURE 1: Differences between MCAL and Active Learning. Active learning outputs an ML model using few samples from the data set. MCAL completely annotates and outputs the dataset. MCAL must also use the ML model to annotate samples reliably (red arrow). core, MCAL learns on-the-fly an accuracy model that, given a number of samples to human-label, and a number to machine label, can predict the overall accuracy of the resultant set of labeled samples. Intuitively, this model implicitly captures the complexity of the classifier and the data set. It also uses a cost model for training and labeling. MCAL proceeds iteratively. At each step, it uses the accuracy model and the cost model to search for the best combination of the number of samples to human-label and to machine-label that would minimize total cost. It obtains the human-labels, trains the classifier D with these, dynamically updates the accuracy model, and machine-labels un-labeled samples if it determines that additional training cannot further reduce cost. | MCAL: MINIMUM COST HUMAN-MACHINE ACTIVE LABELING |
d255440465 | Skip-Attention: Improving Vision Transformers by Paying Less Attention | |
d244729050 | We present Path Integral Sampler (PIS), a novel algorithm to draw samples from unnormalized probability density functions. The PIS is built on the Schrödinger bridge problem which aims to recover the most likely evolution of a diffusion process given its initial distribution and terminal distribution. The PIS draws samples from the initial distribution and then propagates the samples through the Schrödinger bridge to reach the terminal distribution. Applying the Girsanov theorem, with a simple prior diffusion, we formulate the PIS as a stochastic optimal control problem whose running cost is the control energy and terminal cost is chosen according to the target distribution. By modeling the control as a neural network, we establish a sampling algorithm that can be trained end-to-end. We provide theoretical justification of the sampling quality of PIS in terms of Wasserstein distance when sub-optimal control is used. Moreover, the path integrals theory is used to compute importance weights of the samples to compensate for the bias induced by the sub-optimality of the controller and time-discretization. We experimentally demonstrate the advantages of PIS compared with other startof-the-art sampling methods on a variety of tasks. | PATH INTEGRAL SAMPLER: A STOCHASTIC CONTROL APPROACH FOR SAMPLING |
d226221794 | Recently, invariant risk minimization (IRM) was proposed as a promising solution to address out-of-distribution (OOD) generalization. However, it is unclear when IRM should be preferred over the widely-employed empirical risk minimization (ERM) framework. In this work, we analyze both these frameworks from the perspective of sample complexity, thus taking a firm step towards answering this important question. We find that depending on the type of data generation mechanism, the two approaches might have very different finite sample and asymptotic behavior. For example, in the covariate shift setting we see that the two approaches not only arrive at the same asymptotic solution, but also have similar finite sample behavior with no clear winner. For other distribution shifts such as those involving confounders or anti-causal variables, however, the two approaches arrive at different asymptotic solutions where IRM is guaranteed to be close to the desired OOD solutions in the finite sample regime for polynomial generative models, while ERM is biased even asymptotically. We further investigate how different factors -the number of environments, complexity of the model, and IRM penalty weight -impact the sample complexity of IRM in relation to its distance from the OOD solutions. | Empirical or Invariant Risk Minimization? A Sample Complexity Perspective |
d259298560 | Representational straightening refers to a decrease in curvature of visual feature representations of a sequence of frames taken from natural movies. Prior work established straightening in neural representations of the primate primary visual cortex (V1) and perceptual straightening in human behavior as a hallmark of biological vision in contrast to artificial feedforward neural networks which did not demonstrate this phenomenon as they were not explicitly optimized to produce temporally predictable movie representations. Here, we show robustness to noise in the input image can produce representational straightening in feedforward neural networks. Both adversarial training (AT) and base classifiers for Random Smoothing (RS) induced remarkably straightened feature codes. Demonstrating their utility within the domain of natural movies, these codes could be inverted to generate intervening movie frames by linear interpolation in the feature space even though they were not trained on these trajectories. Demonstrating their biological utility, we found that AT and RS training improved predictions of neural data in primate V1 over baseline models providing a parsimonious, bio-plausible mechanism -noise in the sensory input stages -for generating representations in the early visual cortex. Finally, we compared the geometric properties of frame representations in these networks to better understand how they produced representations that mimicked the straightening phenomenon from biology. Overall, this work elucidating emergent properties of robust neural networks demonstrates that it is not necessary to utilize predictive objectives or train directly on natural movie statistics to achieve models supporting straightened movie representations similar to human perception that also predict V1 neural responses. | BRAIN-LIKE REPRESENTATIONAL STRAIGHTENING OF NATURAL MOVIES IN ROBUST FEEDFORWARD NEURAL NETWORKS |
d238583478 | Graph neural networks achieve high accuracy in link prediction by jointly leveraging graph topology and node attributes. Topology, however, is represented indirectly; state-of-the-art methods based on subgraph classification label nodes with distance to the target link, so that, although topological information is present, it is tempered by pooling. This makes it challenging to leverage features like loops and motifs associated with network formation mechanisms. We propose a link prediction algorithm based on a new pooling scheme called WalkPool. WalkPool combines the expressivity of topological heuristics with the feature-learning ability of neural networks. It summarizes a putative link by random walk probabilities of adjacent paths. Instead of extracting transition probabilities from the original graph, it computes the transition matrix of a "predictive" latent graph by applying attention to learned features; this may be interpreted as feature-sensitive topology fingerprinting. WalkPool can leverage unsupervised node features or be combined with GNNs and trained end-to-end. It outperforms state-of-the-art methods on all common link prediction benchmarks, both homophilic and heterophilic, with and without node attributes. Applying WalkPool to a set of unsupervised GNNs significantly improves prediction accuracy, suggesting that it may be used as a general-purpose graph pooling scheme. * These two authors have equal contribution. † To whom correspondence should be addressed. 1 Mathematical topology studies (global) properties of shapes that are preserved under homeomorphisms. Our use of "topology" to refer to local patterns is common in the network literature. | Neural Link Prediction with Walk Pooling |
d67856138 | We consider the dictionary learning problem, where the aim is to model the given data as a linear combination of a few columns of a matrix known as a dictionary, where the sparse weights forming the linear combination are known as coefficients. Since the dictionary and coefficients, parameterizing the linear model are unknown, the corresponding optimization is inherently non-convex. This was a major challenge until recently, when provable algorithms for dictionary learning were proposed. Yet, these provide guarantees only on the recovery of the dictionary, without explicit recovery guarantees on the coefficients. Moreover, any estimation error in the dictionary adversely impacts the ability to successfully localize and estimate the coefficients. This potentially limits the utility of existing provable dictionary learning methods in applications where coefficient recovery is of interest. To this end, we develop NOODL: a simple Neurally plausible alternating Optimization-based Online Dictionary Learning algorithm, which recovers both the dictionary and coefficients exactly at a geometric rate, when initialized appropriately. Our algorithm, NOODL, is also scalable and amenable for large scale distributed implementations in neural architectures, by which we mean that it only involves simple linear and non-linear operations. Finally, we corroborate these theoretical results via experimental evaluation of the proposed algorithm with the current state-of-the-art techniques. Following this, Geng and Wright (2014) and Jenatton et al. (2012) extended these results to the overcomplete case (n < m), and the noisy case, respectively. Concurrently, Jung et al. (2014; 2016) studied the nature of the DL problem for S(·) = · 1 (in (P1)), and derived a lower-bound on the minimax risk of the DL problem. However, these works do not provide any algorithms for DL. Motivated from these theoretical advances, Spielman et al. (2012) proposed an algorithm for the under-complete case n ≥ m that works up-to a sparsity of k = O( √ n). Later, Agarwal et al. (2014) and Arora et al. (2014) proposed clustering-based provable algorithms for the overcomplete setting, motivated from MOD (Engan et al., 1999) and k-SVD (Aharon et al., 2006), respectively. Here, in addition to requiring stringent conditions on dictionary initialization, Agarwal et al. (2014) alternates between solving a quadratic program for coefficients and an MOD-like (Engan et al. , 1999) update for the dictionary, which is too expensive in practice. Recently, a DL algorithm that works for almost linear sparsity was proposed byBarak et al. (2015); however, as shown inTable 1, this algorithm may result in exponential running time. Finally, Arora et al.(2015)proposed a provable online DL algorithm, which provided improvements on initialization, sparsity, and sample complexity, and is closely related to our work. We summarize the relevant provable DL techniques inTable 1.The algorithms discussed above implicitly assume that the coefficients can be recovered, after dictionary recovery, via some sparse approximation technique. However, as alluded to earlier, the guarantees for coefficient recovery -when the dictionary is known approximately -may be limited to some 2 norm bounds(Donoho et al., 2006). This means that, the resulting coefficient estimates may not even be sparse. Therefore, for practical applications, there is a need for efficient online algorithms with guarantees, which serves as the primary motivation for our work.Notation. Given an integer n, we denote [n] = {1, 2, . . . , n}. The bold upper-case and lowercase letters are used to denote matrices M and vectors v, respectively. M i , M (i,:) , M ij , and v i (and v(i)) denote the i-th column, i-th row, (i, j) element of a matrix, and i-th element of a vector, respectively. The superscript (·) (n) denotes the n-th iterate, while the subscript (·) (n) is reserved for the n-th data sample. Given a matrix M, we use M and M F as the spectral norm and Frobenius norm. Given a vector v, we use v , v 0 , and v 1 to denote the 2 norm, 0 (number of nonzero entries), and 1 norm, respectively. We also use standard notations O(·), Ω(·) ( O(·), Ω(·)) to indicate the asymptotic behavior (ignoring logarithmic factors). Further, we use g(n) = O * (f (n)) to indicate that g(n) ≤ Lf (n) for a small enough constant L, which is independent of n. We use c(·) for constants parameterized by the quantities in (·). T τ (z) := z · 1 |z|≥τ denotes the hardthresholding operator, where "1" is the indicator function. We use supp(·) for the support (the set of non-zero elements) and sign(·) for the element-wise sign.ALGORITHMWe now detail the specifics of our algorithm -NOODL, outlined in Algorithm 1. NOODL recovers both the dictionary and the coefficients exactly given an appropriate initial estimate A (0) of the dictionary. Specifically, it requires A (0) to be ( 0 , 2)-close to A * for 0 = O * (1/ log(n)), where ( , κ)-closeness is defined as follows. This implies that, the initial dictionary estimate needs to be column-wise, and in spectral norm sense, close to A * , which can be achieved via certain initialization algorithms, such as those presented inArora et al. (2015). Definition 1 (( , κ)-closeness). A dictionary A is ( , κ)-close to A * if A − A * ≤ κ A * , and if there is a permutation π : [m] → [m] and a collection of signs σ : [m] → {±1} such that σ(i)A π(i) − A * i ≤ , ∀ i ∈ [m]. | NOODL: PROVABLE ONLINE DICTIONARY LEARNING AND SPARSE CODING |
d258048507 | Meta-learning has arisen as a successful method for improving training performance by training over many similar tasks, especially with deep neural networks (DNNs). However, the theoretical understanding of when and why overparameterized models such as DNNs can generalize well in meta-learning is still limited. As an initial step towards addressing this challenge, this paper studies the generalization performance of overfitted meta-learning under a linear regression model with Gaussian features. In contrast to a few recent studies along the same line, our framework allows the number of model parameters to be arbitrarily larger than the number of features in the ground truth signal, and hence naturally captures the overparameterized regime in practical deep meta-learning. We show that the overfitted min ℓ2-norm solution of model-agnostic meta-learning (MAML) can be beneficial, which is similar to the recent remarkable findings on "benign overfitting" and "double descent" phenomenon in the classical (single-task) linear regression. However, due to the uniqueness of meta-learning such as task-specific gradient descent inner training and the diversity/fluctuation of the ground-truth signals among training tasks, we find new and interesting properties that do not exist in single-task linear regression. We first provide a highprobability upper bound (under reasonable tightness) on the generalization error, where certain terms decrease when the number of features increases. Our analysis suggests that benign overfitting is more significant and easier to observe when the noise and the diversity/fluctuation of the ground truth of each training task are large. Under this circumstance, we show that the overfitted min ℓ2-norm solution can achieve an even lower generalization error than the underparameterized solution. | Theoretical Characterization of the Generalization Performance of Overfitted Meta-Learning |
d263830943 | This work considers a rather general and broad class of Markov chains, Ito chains that look like Euler-Maryama discretization of some Stochastic Differential Equation.The chain we study is a unified framework for theoretical analysis.It comes with almost arbitrary isotropic and state-dependent noise instead of normal and state-independent one, as in most related papers.Moreover, our chain's drift and diffusion coefficient can be inexact to cover a wide range of applications such as Stochastic Gradient Langevin Dynamics, sampling, Stochastic Gradient Descent, or Stochastic Gradient Boosting.We prove an upper bound for W 2 -distance between laws of the Ito chain and the corresponding Stochastic Differential Equation.These results improve or cover most of the known estimates.Moreover, for some particular cases, our analysis is the first. | Ito Diffusion Approximation of Universal Ito Chains for Sampling, Optimization and Boosting |
d23873820 | We explore efficient neural architecture search methods and present a simple yet powerful evolutionary algorithm that can discover new architectures achieving state of the art results. Our approach combines a novel hierarchical genetic representation scheme that imitates the modularized design pattern commonly adopted by human experts, and an expressive search space that supports complex topologies. Our algorithm efficiently discovers architectures that outperform a large number of manually designed models for image classification, obtaining top-1 error of 3.6% on CIFAR-10 and 20.3% when transferred to ImageNet, which is competitive with the best existing neural architecture search approaches and represents the new state of the art for evolutionary strategies on this task. We also present results using random search, achieving 0.3% less top-1 accuracy on CIFAR-10 and 0.1% less on ImageNet whilst reducing the architecture search time from 36 hours down to 1 hour. * Work done at DeepMind. | HIERARCHICAL REPRESENTATIONS FOR EFFICIENT ARCHITECTURE SEARCH |
d235614244 | As an essential ingredient of modern deep learning, attention mechanism, especially self-attention, plays a vital role in the global correlation discovery. However, is hand-crafted attention irreplaceable when modeling the global context? Our intriguing finding is that self-attention is not better than the matrix decomposition (MD) model developed 20 years ago regarding the performance and computational cost for encoding the long-distance dependencies. We model the global context issue as a low-rank recovery problem and show that its optimization algorithms can help design global information blocks. This paper then proposes a series of Hamburgers, in which we employ the optimization algorithms for solving MDs to factorize the input representations into sub-matrices and reconstruct a low-rank embedding. Hamburgers with different MDs can perform favorably against the popular global context module self-attention when carefully coping with gradients back-propagated through MDs. Comprehensive experiments are conducted in the vision tasks where it is crucial to learn the global context, including semantic segmentation and image generation, demonstrating significant improvements over self-attention and its variants. Code is available. | IS ATTENTION BETTER THAN MATRIX DECOMPOSITION? |
d222132819 | Role-based learning holds the promise of achieving scalable multi-agent learning by decomposing complex tasks using roles. However, it is largely unclear how to efficiently discover such a set of roles. To solve this problem, we propose to first decompose joint action spaces into restricted role action spaces by clustering actions according to their effects on the environment and other agents. Learning a role selector based on action effects makes role discovery much easier because it forms a bi-level learning hierarchy -the role selector searches in a smaller role space and at a lower temporal resolution, while role policies learn in significantly reduced primitive action-observation spaces. We further integrate information about action effects into the role policies to boost learning efficiency and policy generalization. By virtue of these advances, our method (1) outperforms the current state-of-the-art MARL algorithms on 10 of the 14 scenarios that comprise the challenging StarCraft II micromanagement benchmark and (2) achieves rapid transfer to new environments with three times the number of agents. Demonstrative videos are available at https:// sites.google.com/ view/ rode-marl.Eamonn Butler. The condensed wealth of nations. Centre for Independent Studies, 2012. | RODE: LEARNING ROLES TO DECOMPOSE MULTI-AGENT TASKS |
d222177039 | In this paper, we revisit variational intrinsic control (VIC), an unsupervised reinforcement learning method for finding the largest set of intrinsic options available to an agent. In the original work byGregor et al. (2016), two VIC algorithms were proposed: one that represents the options explicitly, and the other that does it implicitly. We show that the intrinsic reward used in the latter is subject to bias in stochastic environments, causing convergence to suboptimal solutions. To correct this behavior, we propose two methods respectively based on the transitional probability model and Gaussian Mixture Model. We substantiate our claims through rigorous mathematical derivations and experimental analyses. | VARIATIONAL INTRINSIC CONTROL REVISITED |
d247939726 | The goal of dynamic scene deblurring is to remove the motion blur in a given image. Typical learning-based approaches implement their solutions by minimizing the L1 or L2 distance between the output and the reference sharp image. Recent attempts adopt visual recognition features in training to improve the perceptual quality. However, those features are primarily designed to capture high-level contexts rather than low-level structures such as blurriness. Instead, we propose a more direct way to make images sharper by exploiting the inverse task of deblurring, namely, reblurring. Reblurring amplifies the remaining blur to rebuild the original blur, however, a well-deblurred clean image with zero-magnitude blur is hard to reblur. Thus, we design two types of reblurring loss functions for better deblurring. The supervised reblurring loss at training stage compares the amplified blur between the deblurred and the sharp images. The self-supervised reblurring loss at inference stage inspects if there noticeable blur remains in the deblurred. Our experimental results on large-scale benchmarks and real images demonstrate the effectiveness of the reblurring losses in improving the perceptual quality of the deblurred images in terms of NIQE and LPIPS scores as well as visual sharpness. , et al. Image quality assessment: from error visibility to structural similarity. IEEE TIP, 13(4):600-612, 2004. . The unreasonable effectiveness of deep features as a perceptual metric. In CVPR, 2018b. | CLEAN IMAGES ARE HARD TO REBLUR: EXPLOITING THE ILL-POSED INVERSE TASK FOR DYNAMIC SCENE DEBLURRING |
d264825424 | Text-to-3D generation has made remarkable progress recently, particularly with methods based on Score Distillation Sampling (SDS) that leverages pre-trained 2D diffusion models.While the usage of classifier-free guidance is well acknowledged to be crucial for successful optimization, it is considered an auxiliary trick rather than the most essential component.In this paper, we re-evaluate the role of classifier-free guidance in score distillation and discover a surprising finding: the guidance alone is enough for effective text-to-3D generation tasks.We name this method Classifier Score Distillation (CSD), which can be interpreted as using an implicit classification model for generation.This new perspective reveals new insights for understanding existing techniques.We validate the effectiveness of CSD across a variety of text-to-3D tasks including shape generation, texture synthesis, and shape editing, achieving results superior to those of state-of-the-art methods.Our project page is https://xinyu-andy.github.io/Classifier-Score-Distillation | TEXT-TO-3D WITH CLASSIFIER SCORE DISTILLATION |
d263791008 | In Causal Bayesian Optimization (CBO), an agent intervenes on an unknown structural causal model to maximize a downstream reward variable. In this paper, we consider the generalization where other agents or external events also intervene on the system, which is key for enabling adaptiveness to non-stationarities such as weather changes, market forces, or adversaries. We formalize this generalization of CBO as Adversarial Causal Bayesian Optimization (ACBO) and introduce the first algorithm for ACBO with bounded regret: Causal Bayesian Optimization with Multiplicative Weights (CBO-MW). Our approach combines a classical online learning strategy with causal modeling of the rewards. To achieve this, it computes optimistic counterfactual reward estimates by propagating uncertainty through the causal graph. We derive regret bounds for CBO-MW that naturally depend on graph-related quantities. We further propose a scalable implementation for the case of combinatorial interventions and submodular rewards. Empirically, CBO-MW outperforms non-causal and non-adversarial Bayesian optimization methods on synthetic environments and environments based on real-word data. Our experiments include a realistic demonstration of how CBO-MW can be used to learn users' demand patterns in a shared mobility system and reposition vehicles in strategic areas. | ADVERSARIAL CAUSAL BAYESIAN OPTIMIZATION |
d253080708 | Graph neural networks (GNNs) are able to leverage the structure of graph data by passing messages along the edges of the graph. While this allows GNNs to learn features depending on the graph structure, for certain graph topologies it leads to inefficient information propagation and a problem known as oversquashing. This has recently been linked with the curvature and spectral gap of the graph. On the other hand, adding edges to the message-passing graph can lead to increasingly similar node representations and a problem known as oversmoothing. We propose a computationally efficient algorithm that prevents oversquashing by systematically adding edges to the graph based on spectral expansion. We combine this with a relational architecture, which lets the GNN preserve the original graph structure and provably prevents oversmoothing. We find experimentally that our algorithm outperforms existing graph rewiring methods in several graph classification tasks. | FOSR: FIRST-ORDER SPECTRAL REWIRING FOR ADDRESSING OVERSQUASHING IN GNNS |
d53113561 | Conversational machine comprehension requires a deep understanding of the conversation history. To enable traditional, single-turn models to encode the history comprehensively, we introduce FLOW, a mechanism that can incorporate intermediate representations generated during the process of answering previous questions, through an alternating parallel processing structure. Compared to shallow approaches that concatenate previous questions/answers as input, FLOW integrates the latent semantics of the conversation history more deeply. Our model, FLOWQA, shows superior performance on two recently proposed conversational challenges (+7.2% F 1 on CoQA and +4.0% on QuAC). The effectiveness of FLOW also shows in other tasks. By reducing sequential instruction understanding to conversational machine comprehension, FLOWQA outperforms the best models on all three domains in SCONE, with +1.8% to +4.4% improvement in accuracy. * Work done during internship at Allen Institute of Artificial Intelligence. 1 Detailed explanation of existing models is in Section 4. 2 We use "reasoning" to refer to the model process of finding the answer. | FLOWQA: GRASPING FLOW IN HISTORY FOR CONVERSATIONAL MACHINE COMPREHENSION |
d147704157 | The Adam algorithm has become extremely popular for large-scale machine learning. Under convexity condition, it has been proved to enjoy a data-dependant O( √ T ) regret bound where T is the time horizon. However, whether strong convexity can be utilized to further improve the performance remains an open problem. In this paper, we give an affirmative answer by developing a variant of Adam (referred to as SAdam) which achieves a data-dependant O(log T ) regret bound for strongly convex functions. The essential idea is to maintain a faster decaying yet under controlled step size for exploiting strong convexity. In addition, under a special configuration of hyperparameters, our SAdam reduces to SC-RMSprop, a recently proposed variant of RMSprop for strongly convex functions, for which we provide the first data-dependent logarithmic regret bound. Empirical results on optimizing strongly convex functions and training deep networks demonstrate the effectiveness of our method.Preprint. Under review. | SAdam: A Variant of Adam for Strongly Convex Functions |
d229923128 | Knowledge distillation (KD) is essential for training non-autoregressive translation (NAT) models by reducing the complexity of the raw data with an autoregressive teacher model. In this study, we empirically show that as a side effect of this training, the lexical choice errors on low-frequency words are propagated to the NAT model from the teacher model. To alleviate this problem, we propose to expose the raw data to NAT models to restore the useful information of low-frequency words, which are missed in the distilled data. To this end, we introduce an extra Kullback-Leibler divergence term derived by comparing the lexical choice of NAT model and that embedded in the raw data. Experimental results across language pairs and model architectures demonstrate the effectiveness and universality of the proposed approach. Extensive analyses confirm our claim that our approach improves performance by reducing the lexical choice errors on low-frequency words. Encouragingly, our approach pushes the SOTA NAT performance on the WMT14 English-German and WMT16 Romanian-English datasets up to 27.8 and 33.8 BLEU points, respectively. The source code will be released. | UNDERSTANDING AND IMPROVING LEXICAL CHOICE IN NON-AUTOREGRESSIVE TRANSLATION |
d238856778 | Few-shot class incremental learning-the problem of updating a trained classifier to discriminate among an expanded set of classes with limited labeled data-is a key challenge for machine learning systems deployed in non-stationary environments. Existing approaches to the problem rely on complex model architectures and training procedures that are difficult to tune and re-use. In this paper, we present an extremely simple approach that enables the use of ordinary logistic regression classifiers for few-shot incremental learning. The key to this approach is a new family of subspace regularization schemes that encourage weight vectors for new classes to lie close to the subspace spanned by the weights of existing classes. When combined with pretrained convolutional feature extractors, logistic regression models trained with subspace regularization outperform specialized, state-of-the-art approaches to few-shot incremental image classification by up to 22% on the miniImageNet dataset. Because of its simplicity, subspace regularization can be straightforwardly extended to incorporate additional background information about the new classes (including class names and descriptions specified in natural language); these further improve accuracy by up to 2%. Our results show that simple geometric regularization of class representations offers an effective tool for continual learning. 1 1 Code for the experiments is released under https://github.com/feyzaakyurek/subspace-reg. | SUBSPACE REGULARIZERS FOR FEW-SHOT CLASS INCREMENTAL LEARNING |
d235489741 | We advocate for a practical Maximum Likelihood Estimation (MLE) approach towards designing loss functions for regression and forecasting, as an alternative to the typical approach of direct empirical risk minimization on a specific target metric. The MLE approach is better suited to capture inductive biases such as prior domain knowledge in datasets, and can output post-hoc estimators at inference time that can optimize different types of target metrics. We present theoretical results to demonstrate that our approach is competitive with any estimator for the target metric under some general conditions. In two example practical settings, Poisson and Pareto regression, we show that our competitive results can be used to prove that the MLE approach has better excess risk bounds than directly minimizing the target metric. We also demonstrate empirically that our method instantiated with a well-designed general purpose mixture likelihood family can obtain superior performance for a variety of tasks across time-series forecasting and regression datasets with different data distributions. | On the benefits of maximum likelihood estimation for Regression and Forecasting |
d263672002 | Given a set of K probability densities, we consider the multimarginal generative modeling problem of learning a joint distribution that recovers these densities as marginals.The structure of this joint distribution should identify multi-way correspondences among the prescribed marginals.We formalize an approach to this task within a generalization of the stochastic interpolant framework, leading to efficient learning algorithms built upon dynamical transport of measure.Our generative models are defined by velocity and score fields that can be characterized as the minimizers of simple quadratic objectives, and they are defined on a simplex that generalizes the time variable in the usual dynamical transport framework.The resulting transport on the simplex is influenced by all marginals, and we show that multi-way correspondences can be extracted.The identification of such correspondences has applications to style transfer, algorithmic fairness, and data decorruption.In addition, the multimarginal perspective enables an efficient algorithm for reducing the dynamical transport cost in the ordinary two-marginal setting.We demonstrate these capacities with several numerical examples. | MULTIMARGINAL GENERATIVE MODELING WITH STOCHASTIC INTERPOLANTS |
d260351224 | Assigning importance weights to adversarial data has achieved great success in training adversarially robust networks under limited model capacity. However, existing instance-reweighted adversarial training (AT) methods heavily depend on heuristics and/or geometric interpretations to determine those importance weights, making these algorithms lack rigorous theoretical justification/guarantee. Moreover, recent research has shown that adversarial training suffers from a severe non-uniform robust performance across the training distribution, e.g., data points belonging to some classes can be much more vulnerable to adversarial attacks than others. To address both issues, in this paper, we propose a novel doublyrobust instance reweighted AT framework, which allows to obtain the importance weights via exploring distributionally robust optimization (DRO) techniques, and at the same time boosts the robustness on the most vulnerable examples. In particular, our importance weights are obtained by optimizing the KL-divergence regularized loss function, which allows us to devise new algorithms with a theoretical convergence guarantee. Experiments on standard classification datasets demonstrate that our proposed approach outperforms related state-of-the-art baseline methods in terms of average robust performance, and at the same time improves the robustness against attacks on the weakest data points. Codes will be available soon. arXiv:2308.00311v1 [cs.LG] 1 Aug 2023 data consumes much more model capacity compared to their clean counterpart, making the minimization of the uniform average adversarial loss a very pessimistic goal, as argued in Zhang et al. (2020b). To overcome this limitation, recent works Zhang et al. (2020b); Liu et al. (2021a); Zeng et al. (2021); Ding et al. (2018)assign an importance weight to each data point in the training distribution, in order to emphasize the ones that are critical to determining the model's decision boundaries. By allowing more careful exploitation of the limited model capacity, such a simple instance-reweighted scheme combined with traditional adversarial training has yielded a significant boost in the robust performance of current adversarially trained models. Yet, existing methods for instance-reweighted AT mostly adopt heuristic techniques and/or geometric intuitions in order to compute the instance weights, which makes these algorithms lack a principled and rigorous theoretical justification/guarantee. This hence motivates the following question we ask:How to systematically determine the importance weights via a principled approach, rather than resorting to heuristics/interpretations which are often sub-optimal? Moreover, as observed inTian et al. (2021), another critical limitation of the transitional AT method is that it suffers a severe non-uniform performance across the empirical distribution. For example, while the average robust performance of the AT method on the CIFAR10 dataset can be as high as 49%, the robust accuracy for the weakest class is as low as 14%, which depicts a huge disparity in robust performance across different classes. We note that such a non-uniform performance across classes is also slightly observed in the standard training with clean data, but its severity is much worsened in adversarial training (seeFigure 1). Indeed, this is a critical limitation that requires special attention as, in a real-world situation, a more intelligent attacker can, in fact, decide which examples to attack so as to achieve a much higher success rate (e.g., 87% when attacking the most vulnerable class). This non-uniform robust performance is even worsened in the case of imbalanced training distributions Wu et al.(2021);Wang et al. (2022), where the robust performance for the most vulnerable class can be as low as 0%. This motivates our second question given below:Can such an issue of non-uniform performance particularly over imbalanced datasets be addressed at the instance level simultaneously as we design the importance weights to address the first question? In this paper, we propose a novel doubly robust instance reweighted optimization approach to address both of the above questions. | Doubly Robust Instance-Reweighted Adversarial Training |
d108297416 | This paper describes a new form of unsupervised learning, whose input is a set of unlabeled points that are assumed to be local maxima of an unknown value function v in an unknown subset of the vector space. Two functions are learned: (i) a set indicator c, which is a binary classifier, and (ii) a comparator function h that given two nearby samples, predicts which sample has the higher value of the unknown function v. Loss terms are used to ensure that all training samples x are a local maxima of v, according to h and satisfy c(x) = 1. Therefore, c and h provide training signals to each other: a point x in the vicinity of x satisfies c(x) = −1 or is deemed by h to be lower in value than x. We present an algorithm, show an example where it is more efficient to use local maxima as an indicator function than to employ conventional classification, and derive a suitable generalization bound. Our experiments show that the method is able to outperform one-class classification algorithms in the task of anomaly detection and also provide an additional signal that is extracted in a completely unsupervised way.In another instance, consider the weights of multiple neural networks trained to minimize the same loss on the same training data, each using a different random initialization. Considering the weights of each trained neural network as a single vector in a sample domain, also fits into the framework of local optimality under constraints. By the nature of the problem, the obtained weights are the local optimum of some loss optimization process. In addition, the weights are sometimes subject to constraints, e.g., by using weight normalization.The task tackled in this paper is learning the value function and the constraints, by observing only the local maxima of the value function among points that satisfy the constraints. This is an unsupervised problem: no labels are given in addition to the samples.The closest computational problem in the literature is one-class classification(Moya et al., 1993), where one learns a classifier c in order to model a set of unlabeled training samples, all from a single class. In our formulation, two functions are learned: a classifier and a separate value function. Splitting the modeling task between the two, a simpler classifier can be used (we prove this for a specific case) and we also observe improved empirical performance. In addition, we show that the value function, which is trained with different losses and structure from those of the classifier, models a different aspect of the training set. For example, if the samples are images from a certain class, the classifier would capture class membership and the value function would encode image quality. The emergence of a quality model makes sense, since the training images are often homogeneous in their quality. | UNSUPERVISED LEARNING OF THE SET OF LOCAL MAXIMA |
d256663850 | More and more evidence has shown that strengthening layer interactions can enhance the representation power of a deep neural network, while self-attention excels at learning interdependencies by retrieving query-activated information. Motivated by this, we devise a cross-layer attention mechanism, called multi-head recurrent layer attention (MRLA), that sends a query representation of the current layer to all previous layers to retrieve query-related information from different levels of receptive fields. A light-weighted version of MRLA is also proposed to reduce the quadratic computation cost. The proposed layer attention mechanism can enrich the representation power of many state-of-the-art vision networks, including CNNs and vision transformers. Its effectiveness has been extensively evaluated in image classification, object detection and instance segmentation tasks, where improvements can be consistently observed. For example, our MRLA can improve 1.6% Top-1 accuracy on ResNet-50, while only introducing 0.16M parameters and 0.07B FLOPs. Surprisingly, it can boost the performances by a large margin of 3-4% box AP and mask AP in dense prediction tasks. Our code is available at https://github.com | CROSS-LAYER RETROSPECTIVE RETRIEVING VIA LAYER ATTENTION |
d263829757 | We introduce JointNet, a novel neural network architecture for modeling the joint distribution of images and an additional dense modality (e.g., depth maps).Joint-Net is extended from a pre-trained text-to-image diffusion model, where a copy of the original network is created for the new dense modality branch and is densely connected with the RGB branch.The RGB branch is locked during network finetuning, which enables efficient learning of the new modality distribution while maintaining the strong generalization ability of the large-scale pre-trained diffusion model.We demonstrate the effectiveness of JointNet by using RGBD diffusion as an example and through extensive experiments, showcasing its applicability in a variety of applications, including joint RGBD generation, dense depth prediction, depth-conditioned image generation, and coherent tile-based 3D panorama generation. | JOINTNET: EXTENDING TEXT-TO-IMAGE DIFFUSION FOR DENSE DISTRIBUTION MODELING |
d263136146 | Recent advancements in autonomous driving have relied on data-driven approaches, which are widely adopted but face challenges including dataset bias, overfitting, and uninterpretability.Drawing inspiration from the knowledge-driven nature of human driving, we explore the question of how to instill similar capabilities into autonomous driving systems and summarize a paradigm that integrates an interactive environment, a driver agent, as well as a memory component to address this question.Leveraging large language models (LLMs) with emergent abilities, we propose the DiLu framework, which combines a Reasoning and a Reflection module to enable the system to perform decision-making based on common-sense knowledge and evolve continuously.Extensive experiments prove DiLu's capability to accumulate experience and demonstrate a significant advantage in generalization ability over reinforcement learning-based methods.Moreover, DiLu is able to directly acquire experiences from real-world datasets which highlights its potential to be deployed on practical autonomous driving systems.To the best of our knowledge, we are the first to instill knowledge-driven capability into autonomous driving systems from the perspective of how humans drive. | DILU: A KNOWLEDGE-DRIVEN APPROACH TO AUTONOMOUS DRIVING WITH LARGE LANGUAGE MODELS |
d212414027 | We propose an energy-based model (EBM) of protein conformations that operates at atomic scale. The model is trained solely on crystallized protein data. By contrast, existing approaches for scoring conformations use energy functions that incorporate knowledge of physical principles and features that are the complex product of several decades of research and tuning. To evaluate the model, we benchmark on the rotamer recovery task, the problem of predicting the conformation of a side chain from its context within a protein structure, which has been used to evaluate energy functions for protein design. The model achieves performance close to that of the Rosetta energy function, a state-of-the-art method widely used in protein structure prediction and design. An investigation of the models outputs and hidden representations finds that it captures physicochemical properties relevant to protein energy. * Work performed during an internship at Facebook 1 Data and code for experiments are available at https://github.com/facebookresearch/ protein-ebm | ENERGY-BASED MODELS FOR ATOMIC-RESOLUTION PROTEIN CONFORMATIONS |
d259203325 | It is often advantageous to train models on a subset of the available train examples, because the examples are of variable quality or because one would like to train with fewer examples, without sacrificing performance. We present Gradient Information Optimization (GIO), a scalable, task-agnostic approach to this data selection problem that requires only a small set of (unlabeled) examples representing a target distribution. GIO begins from a natural, information-theoretic objective that is intractable in practice. Our contribution is in showing that it can be made highly scalable through a simple relaxation of the objective and a highly efficient implementation. In experiments with machine translation, spelling correction, and image recognition, we show that GIO delivers outstanding results with very small train sets. These findings are robust to different representation models and hyperparameters for GIO itself. GIO is task-and domain-agnostic and can be applied out-of-the-box to new datasets and domains.Active learning. Active learning methods (e.g. Sener and Savarese, 2018; Gal et al., 2017; Kirsch et al., 2019) can be cast as data selection methods in our sense. In active learning, one iteratively chooses new unlabeled training examples to label, with the goal of efficiently creating a powerful train set. By contrast, GIO makes no use of labels and is oriented towards the goal of identifying a subset of existing cases to use for training. Additionally, active learning is most suited to classification problems, whereas GIO works with any arbitrary task.Heuristic. GIO is closer to recent methods in which one uses a large language model to generate a large number of candidate texts and then extracts a subset of them based on a specific criteria. For example, Brown et al. (2020) develop a heuristic method to filter CommonCrawl based on a trained classifier's probability that datapoints are high quality. Similarly, Wenzek et al. (2020) develop a pipeline to clean CommonCrawl based principally on the perplexity of an LM trained on high quality text, and Xie et al. (2023) develop a sampling technique based on approximate n-gram counts.Like GIO, these heuristic methods aim to select a subset of data that is higher quality and more relevant. However, they are either highly tailored to their particular tasks or they require very large numbers of examples (to develop classifiers or construct target probabilities). By contrast, GIO is task-and domain-agnostic, it can be applied plug-and-play to a new task and dataset, and it requires comparatively few gold examples X to serve as the target distribution.Similarity Search. Methods using vector or n-gram similarity search can also be used for data selection at scale (e.g. Johnson et al., 2017;Bernhardsson, 2017;Santhanam et al., 2022). The technique would index G and X and retrieve the top-k datapoints from G for each point in X. Like our method, similarity search works in a continuous space. However, similarity search can be prone to selecting suboptimal points; we review such a case in detail in Section 3.4. Additionally, similarity search does not have a natural stopping criterion and requires data size to be chosen arbitrarily. Is 10% data enough? 20%? We don't know a priori. And if the data in G is far away from X, similarity search will still choose it up to the desired data size. Recently, Yao et al. (2022) use a BM25 retrieval method for data selection, with strong results. However, BM25 operates on a bag-of-words model, which can make it challenging when the target set is small, and like any similarity search, requires data size to be chosen arbitrarily beforehand. Further, this method only applies to text tasks, whereas GIO applies to any task with continuous representation. | GIO: GRADIENT INFORMATION OPTIMIZATION FOR TRAINING DATASET SELECTION |
d222140947 | Contrastive learning has been adopted as a core method for unsupervised visual representation learning. Without human annotation, the common practice is to perform an instance discrimination task: Given a query image crop, this task labels crops from the same image as positives, and crops from other randomly sampled images as negatives. An important limitation of this label assignment strategy is that it can not reflect the heterogeneous similarity between the query crop and each crop from other images, taking them as equally negative, while some of them may even belong to the same semantic class as the query. To address this issue, inspired by consistency regularization in semi-supervised learning on unlabeled data, we propose Consistent Contrast (CO2), which introduces a consistency regularization term into the current contrastive learning framework. Regarding the similarity of the query crop to each crop from other images as "unlabeled", the consistency term takes the corresponding similarity of a positive crop as a pseudo label, and encourages consistency between these two similarities. Empirically, CO2 improves Momentum Contrast (MoCo) by 2.9% top-1 accuracy on ImageNet linear protocol, 3.8% and 1.1% top-5 accuracy on 1% and 10% labeled semi-supervised settings. It also transfers to image classification, object detection, and semantic segmentation on PASCAL VOC. This shows that CO2 learns better visual representations for these downstream tasks. | CO2: CONSISTENT CONTRAST FOR UNSUPERVISED VISUAL REPRESENTATION LEARNING |
d235485093 | Recent work in equivariant deep learning bears strong similarities to physics. Fields over a base space are fundamental entities in both subjects, as are equivariant maps between these fields. In deep learning, however, these maps are usually defined by convolutions with a kernel, whereas they are partial differential operators (PDOs) in physics. Developing the theory of equivariant PDOs in the context of deep learning could bring these subjects even closer together and lead to a stronger flow of ideas. In this work, we derive a G-steerability constraint that completely characterizes when a PDO between feature vector fields is equivariant, for arbitrary symmetry groups G. We then fully solve this constraint for several important groups. We use our solutions as equivariant drop-in replacements for convolutional layers and benchmark them in that role. Finally, we develop a framework for equivariant maps based on Schwartz distributions that unifies classical convolutions and differential operators and gives insight about the relation between the two. Figure 1: A vector field (left) can be mapped to a scalar field (right) by applying certain partial differential operators (PDOs), such as the Laplacian of the divergence and the 2D curl. Such a PDO from a 2D vector to a scalar field can be represented as a 2 × 1 matrix, where each of the two entries is a one-dimensional PDO that acts on one of the two components of the vector field. Similarly, matrices of PDOs with different dimensions map between other types of fields. Our goal is to find all PDOs for which this map becomes equivariant, for arbitrary types of fields. For the implementation, we will later discretize PDOs as stencils (middle). | STEERABLE PARTIAL DIFFERENTIAL OPERATORS FOR EQUIVARIANT NEURAL NETWORKS |
d264452015 | We present DreamCraft3D, a hierarchical 3D content generation method that produces high-fidelity and coherent 3D objects.We tackle the problem by leveraging a 2D reference image to guide the stages of geometry sculpting and texture boosting.A central focus of this work is to address the consistency issue that existing works encounter.To sculpt geometries that render coherently, we perform score distillation sampling via a view-dependent diffusion model.This 3D prior, alongside several training strategies, prioritizes the geometry consistency but compromises the texture fidelity.We further propose Bootstrapped Score Distillation to specifically boost the texture.We train a personalized diffusion model, Dreambooth, on the augmented renderings of the scene, imbuing it with 3D knowledge of the scene being optimized.The score distillation from this 3D-aware diffusion prior provides view-consistent guidance for the scene.Notably, through an alternating optimization of the diffusion prior and 3D scene representation, we achieve mutually reinforcing improvements: the optimized 3D scene aids in training the scene-specific diffusion model, which offers increasingly view-consistent guidance for 3D optimization.The optimization is thus bootstrapped and leads to substantial texture boosting.With tailored 3D priors throughout the hierarchical generation, DreamCraft3D generates coherent 3D objects with photorealistic renderings, advancing the state-of-the-art in 3D content generation.Code available at https://github.com/deepseek-ai/DreamCraft3D. | DREAMCRAFT3D: HIERARCHICAL 3D GENERATION WITH BOOTSTRAPPED DIFFUSION PRIOR |
d259138847 | This paper studies the curious phenomenon for machine learning models with Transformer architectures that their activation maps are sparse. By activation map we refer to the intermediate output of the multi-layer perceptrons (MLPs) after a ReLU activation function, and by "sparse" we mean that on average very few entries (e.g., 3.0% for T5-Base and 6.3% for ViT-B16) are nonzero for each input to MLP. Moreover, larger Transformers with more layers and wider MLP hidden dimensions are sparser as measured by the percentage of nonzero entries. Through extensive experiments we demonstrate that the emergence of sparsity is a prevalent phenomenon that occurs for both natural language processing and vision tasks, on both training and evaluation data, for Transformers of various configurations, at layers of all depth levels, as well as for other architectures including MLP-mixers and 2-layer MLPs. We show that sparsity also emerges using training datasets with random labels, or with random inputs, or with infinite amount of data, demonstrating that sparsity is not a result of a specific family of datasets. We discuss how sparsity immediately implies a way to significantly reduce the FLOP count and improve efficiency for Transformers. Moreover, we demonstrate perhaps surprisingly that enforcing an even sparser activation via Top-k thresholding with a small value of k brings a collection of desired but missing properties for Transformers, namely less sensitivity to noisy training data, more robustness to input corruptions, and better calibration for their prediction confidence. * Equal contribution arXiv:2210.06313v2 [cs.LG] 9 Jun 2023 | The Lazy Neuron Phenomenon: On Emergence of Activation Sparsity in Transformers |
d236087421 | An Axial Shifted MLP architecture (AS-MLP) is proposed in this paper. Different from MLP-Mixer, where the global spatial feature is encoded for information flow through matrix transposition and one token-mixing MLP, we pay more attention to the local features interaction. By axially shifting channels of the feature map, AS-MLP is able to obtain the information flow from different axial directions, which captures the local dependencies. Such an operation enables us to utilize a pure MLP architecture to achieve the same local receptive field as CNN-like architecture. We can also design the receptive field size and dilation of blocks of AS-MLP, etc, in the same spirit of convolutional neural networks. With the proposed AS-MLP architecture, our model obtains 83.3% Top-1 accuracy with 88M parameters and 15.2 GFLOPs on the ImageNet-1K dataset. Such a simple yet effective architecture outperforms all MLP-based architectures and achieves competitive performance compared to the transformer-based architectures (e.g., Swin Transformer) even with slightly lower FLOPs. In addition, AS-MLP is also the first MLP-based architecture to be applied to the downstream tasks (e.g., object detection and semantic segmentation). The experimental results are also impressive. Our proposed AS-MLP obtains 51.5 mAP on the COCO validation set and 49.5 MS mIoU on the ADE20K dataset, which is competitive compared to the transformer-based architectures. Our AS-MLP establishes a strong baseline of MLP-based architecture.Published as a conference paper at ICLR 2022 local information focuses more on extracting the low-level features. In the transformer-based architectures, Swin Transformer computes the self-attention in a window (7 × 7) instead of the global receptive field, which is similar to directly using a convolution layer with a large kernel size (7×7) to cover the local receptive field. Some other papers have also already emphasized the advantages of local receptive fields, and introduced local information in the transformer, such as Localvit ,NesT (Zhang et al., 2021), etc. Driven by these ideas, we mainly explore the influence of locality on MLP-based architectures.In order to introduce locality into the MLP-based architecture, one of the simplest and most intuitive ideas is to add a window to the MLP-Mixer, and then perform a token-mixing projection of the local features within the window, just as done in Swin Transformer and LINMAPPER (Fang et al., 2021). However, if we divide the window (e.g., 7 × 7) and perform the token-mixing projection in the window, then the linear layer has the 49 × 49 parameters shared between windows, which greatly limits the model capacity and thus affects the learning of parameters and final results. | AS-MLP: AN AXIAL SHIFTED MLP ARCHITECTURE FOR VISION |
d240070408 | Deep neural networks (DNNs) are proved to be vulnerable against backdoor attacks. A backdoor is often embedded in the target DNNs through injecting a backdoor trigger into training examples, which can cause the target DNNs misclassify an input attached with the backdoor trigger. Existing backdoor detection methods often require the access to the original poisoned training data, the parameters of the target DNNs, or the predictive confidence for each given input, which are impractical in many real-world applications, e.g., on-device deployed DNNs. We address the black-box hard-label backdoor detection problem where the DNN is fully black-box and only its final output label is accessible. We approach this problem from the optimization perspective and show that the objective of backdoor detection is bounded by an adversarial objective. Further theoretical and empirical studies reveal that this adversarial objective leads to a solution with highly skewed distribution; a singularity is often observed in the adversarial map of a backdoorinfected example, which we call the adversarial singularity phenomenon. Based on this observation, we propose the adversarial extreme value analysis (AEVA) to detect backdoors in black-box neural networks. AEVA is based on an extreme value analysis of the adversarial map, computed from the monte-carlo gradient estimation. Evidenced by extensive experiments across multiple popular tasks and backdoor attacks, our approach is shown effective in detecting backdoor attacks under the black-box hard-label scenarios. | AEVA: BLACK-BOX BACKDOOR DETECTION USING ADVERSARIAL EXTREME VALUE ANALYSIS |
d256662465 | Incorporating equivariance to symmetry groups as a constraint during neural network training can improve performance and generalization for tasks exhibiting those symmetries, but such symmetries are often not perfectly nor explicitly present. This motivates algorithmically optimizing the architectural constraints imposed by equivariance. We propose the equivariance relaxation morphism, which preserves functionality while reparameterizing a group equivariant layer to operate with equivariance constraints on a subgroup, as well as the [G]-mixed equivariant layer, which mixes layers constrained to different groups to enable within-layer equivariance optimization. We further present evolutionary and differentiable neural architecture search (NAS) algorithms that utilize these mechanisms respectively for equivariance-aware architectural optimization. Experiments across a variety of datasets show the benefit of dynamically constrained equivariance to find effective architectures with approximate equivariance. | EQUIVARIANCE-AWARE ARCHITECTURAL OPTIMIZATION OF NEURAL NETWORKS |
d246276208 | The ability to learn reward functions plays an important role in enabling the deployment of intelligent agents in the real world. However, comparing reward functions, for example as a means of evaluating reward learning methods, presents a challenge. Reward functions are typically compared by considering the behavior of optimized policies, but this approach conflates deficiencies in the reward function with those of the policy search algorithm used to optimize it. To address this challenge,Gleave et al. (2020)propose the Equivalent-Policy Invariant Comparison (EPIC) distance. EPIC avoids policy optimization, but in doing so requires computing reward values at transitions that may be impossible under the system dynamics. This is problematic for learned reward functions because it entails evaluating them outside of their training distribution, resulting in inaccurate reward values that we show can render EPIC ineffective at comparing rewards. To address this problem, we propose the Dynamics-Aware Reward Distance (DARD), a new reward pseudometric. DARD uses an approximate transition model of the environment to transform reward functions into a form that allows for comparisons that are invariant to reward shaping while only evaluating reward functions on transitions close to their training distribution. Experiments in simulated physical domains demonstrate that DARD enables reliable reward comparisons without policy optimization and is significantly more predictive than baseline methods of downstream policy performance when dealing with learned reward functions. † † | DYNAMICS-AWARE COMPARISON OF LEARNED REWARD FUNCTIONS |
d250048824 | Linear time-invariant state space models (SSM) are a classical model from engineering and statistics, that have recently been shown to be very promising in machine learning through the Structured State Space sequence model (S4). A core component of S4 involves initializing the SSM state matrix to a particular matrix called a HiPPO matrix, which was empirically important for S4's ability to handle long sequences. However, the specific matrix that S4 uses was actually derived in previous work for a particular time-varying dynamical system, and the use of this matrix as a time-invariant SSM had no known mathematical interpretation. Consequently, the theoretical mechanism by which S4 models long-range dependencies actually remains unexplained. We derive a more general and intuitive formulation of the HiPPO framework, which provides a simple mathematical interpretation of S4 as a decomposition onto exponentially-warped Legendre polynomials, explaining its ability to capture long dependencies. Our generalization introduces a theoretically rich class of SSMs that also lets us derive more intuitive S4 variants for other bases such as the Fourier basis, and explains other aspects of training S4, such as how to initialize the important timescale parameter. These insights improve S4's performance to 86% on the Long Range Arena benchmark, with 96% on the most difficult Path-X task. * Equal contribution. | How to Train Your HiPPO: State Space Models with Generalized Orthogonal Basis Projections |
d225103395 | A key factor in the success of deep neural networks is the ability to scale models to improve performance by varying the architecture depth and width. This simple property of neural network design has resulted in highly effective architectures for a variety of tasks. Nevertheless, there is limited understanding of effects of depth and width on the learned representations. In this paper, we study this fundamental question. We begin by investigating how varying depth and width affects model hidden representations, finding a characteristic block structure in the hidden representations of larger capacity (wider or deeper) models. We demonstrate that this block structure arises when model capacity is large relative to the size of the training set, and is indicative of the underlying layers preserving and propagating the dominant principal component of their representations. This discovery has important ramifications for features learned by different models, namely, representations outside the block structure are often similar across architectures with varying widths and depths, but the block structure is unique to each model. We analyze the output predictions of different model architectures, finding that even when the overall accuracy is similar, wide and deep models exhibit distinctive error patterns and variations across classes. * Work done as a member of the Google AI Residency program. | DO WIDE AND DEEP NETWORKS LEARN THE SAME THINGS? UNCOVERING HOW NEURAL NETWORK REPRESENTATIONS VARY WITH WIDTH AND DEPTH |
d20827927 | Disentangled representations, where the higher level data generative factors are reflected in disjoint latent dimensions, offer several benefits such as ease of deriving invariant representations, transferability to other tasks, interpretability, etc. We consider the problem of unsupervised learning of disentangled representations from large pool of unlabeled observations, and propose a variational inference based approach to infer disentangled latent factors. We introduce a regularizer on the expectation of the approximate posterior over observed data that encourages the disentanglement. We evaluate the proposed approach using several quantitative metrics and empirically observe significant gains over existing methods in terms of both disentanglement and data likelihood (reconstruction quality). | Variational Inference of Disentangled Latent Concepts from Unlabeled Observations |
d254044657 | Algorithmic fairness plays an increasingly critical role in machine learning research.Several group fairness notions and algorithms have been proposed.However, the fairness guarantee of existing fair classification methods mainly depends on specific data distributional assumptions, often requiring large sample sizes, and fairness could be violated when there is a modest number of samples, which is often the case in practice.In this paper, we propose FaiREE, a fair classification algorithm that can satisfy group fairness constraints with finite-sample and distribution-free theoretical guarantees.FaiREE can be adapted to satisfy various group fairness notions (e.g., Equality of Opportunity, Equalized Odds, Demographic Parity, etc.) and achieve the optimal accuracy.These theoretical guarantees are further supported by experiments on both synthetic and real data.FaiREE is shown to have favorable performance over state-of-the-art algorithms. | FaiREE: Fair Classification with Finite-Sample and Distribution-Free Guarantee |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.