JohnVitz/NLP_Assignment_3_part_3 · Datasets at Hugging Face

2401.07301

Small Language Model Can Self-correct

Generative Language Models (LMs) such as ChatGPT have exhibited remarkable performance across various downstream tasks. Nevertheless, one of their most prominent drawbacks is generating inaccurate or false information with a confident tone. Previous studies have devised sophisticated pipelines and prompts to induce large LMs to exhibit the capability for self-correction. However, large LMs are explicitly prompted to verify and modify its answers separately rather than completing all steps spontaneously like humans. Moreover, these complex prompts are extremely challenging for small LMs to follow. In this paper, we introduce the \underline{I}ntrinsic \underline{S}elf-\underline{C}orrection (ISC) in generative language models, aiming to correct the initial output of LMs in a self-triggered manner, even for those small LMs with 6 billion parameters. Specifically, we devise a pipeline for constructing self-correction data and propose Partial Answer Masking (PAM), aiming to endow the model with the capability for intrinsic self-correction through fine-tuning. We conduct experiments using LMs with parameters sizes ranging from 6 billion to 13 billion in two tasks, including commonsense reasoning and factual knowledge reasoning. Our experiments demonstrate that the outputs generated using ISC outperform those generated without self-correction. We believe that the output quality of even small LMs can be further improved by empowering them with the ability to intrinsic self-correct.

false

true

false

true

false

421,486

2407.01920

To Forget or Not? Towards Practical Knowledge Unlearning for Large Language Models

Large Language Models (LLMs) trained on extensive corpora inevitably retain sensitive data, such as personal privacy information and copyrighted material. Recent advancements in knowledge unlearning involve updating LLM parameters to erase specific knowledge. However, current unlearning paradigms are mired in vague forgetting boundaries, often erasing knowledge indiscriminately. In this work, we introduce KnowUnDo, a benchmark containing copyrighted content and user privacy domains to evaluate if the unlearning process inadvertently erases essential knowledge. Our findings indicate that existing unlearning methods often suffer from excessive unlearning. To address this, we propose a simple yet effective method, MemFlex, which utilizes gradient information to precisely target and unlearn sensitive parameters. Experimental results show that MemFlex is superior to existing methods in both precise knowledge unlearning and general knowledge retaining of LLMs. Code and dataset are released at https://github.com/zjunlp/KnowUnDo.

false

true

false

true

false

true

false

true

false

true

469,511

2106.00992

NVC-Net: End-to-End Adversarial Voice Conversion

Voice conversion has gained increasing popularity in many applications of speech synthesis. The idea is to change the voice identity from one speaker into another while keeping the linguistic content unchanged. Many voice conversion approaches rely on the use of a vocoder to reconstruct the speech from acoustic features, and as a consequence, the speech quality heavily depends on such a vocoder. In this paper, we propose NVC-Net, an end-to-end adversarial network, which performs voice conversion directly on the raw audio waveform of arbitrary length. By disentangling the speaker identity from the speech content, NVC-Net is able to perform non-parallel traditional many-to-many voice conversion as well as zero-shot voice conversion from a short utterance of an unseen target speaker. Importantly, NVC-Net is non-autoregressive and fully convolutional, achieving fast inference. Our model is capable of producing samples at a rate of more than 3600 kHz on an NVIDIA V100 GPU, being orders of magnitude faster than state-of-the-art methods under the same hardware configurations. Objective and subjective evaluations on non-parallel many-to-many voice conversion tasks show that NVC-Net obtains competitive results with significantly fewer parameters.

false

true

false

true

false

238,340

1401.3840

Grounding FO and FO(ID) with Bounds

Grounding is the task of reducing a first-order theory and finite domain to an equivalent propositional theory. It is used as preprocessing phase in many logic-based reasoning systems. Such systems provide a rich first-order input language to a user and can rely on efficient propositional solvers to perform the actual reasoning. Besides a first-order theory and finite domain, the input for grounders contains in many applications also additional data. By exploiting this data, the size of the grounders output can often be reduced significantly. A common practice to improve the efficiency of a grounder in this context is by manually adding semantically redundant information to the input theory, indicating where and when the grounder should exploit the data. In this paper we present a method to compute and add such redundant information automatically. Our method therefore simplifies the task of writing input theories that can be grounded efficiently by current systems. We first present our method for classical first-order logic (FO) theories. Then we extend it to FO(ID), the extension of FO with inductive definitions, which allows for more concise and comprehensive input theories. We discuss implementation issues and experimentally validate the practical applicability of our method.

false

true

false

true

29,956

1803.05209

Building Sparse Deep Feedforward Networks using Tree Receptive Fields

Sparse connectivity is an important factor behind the success of convolutional neural networks and recurrent neural networks. In this paper, we consider the problem of learning sparse connectivity for feedforward neural networks (FNNs). The key idea is that a unit should be connected to a small number of units at the next level below that are strongly correlated. We use Chow-Liu's algorithm to learn a tree-structured probabilistic model for the units at the current level, use the tree to identify subsets of units that are strongly correlated, and introduce a new unit with receptive field over the subsets. The procedure is repeated on the new units to build multiple layers of hidden units. The resulting model is called a TRF-net. Empirical results show that, when compared to dense FNNs, TRF-net achieves better or comparable classification performance with much fewer parameters and sparser structures. They are also more interpretable.

false

true

false

92,601

2001.00426

Graph Signal Processing -- Part III: Machine Learning on Graphs, from Graph Topology to Applications

Many modern data analytics applications on graphs operate on domains where graph topology is not known a priori, and hence its determination becomes part of the problem definition, rather than serving as prior knowledge which aids the problem solution. Part III of this monograph starts by addressing ways to learn graph topology, from the case where the physics of the problem already suggest a possible topology, through to most general cases where the graph topology is learned from the data. A particular emphasis is on graph topology definition based on the correlation and precision matrices of the observed data, combined with additional prior knowledge and structural conditions, such as the smoothness or sparsity of graph connections. For learning sparse graphs (with small number of edges), the least absolute shrinkage and selection operator, known as LASSO is employed, along with its graph specific variant, graphical LASSO. For completeness, both variants of LASSO are derived in an intuitive way, and explained. An in-depth elaboration of the graph topology learning paradigm is provided through several examples on physically well defined graphs, such as electric circuits, linear heat transfer, social and computer networks, and spring-mass systems. As many graph neural networks (GNN) and convolutional graph networks (GCN) are emerging, we have also reviewed the main trends in GNNs and GCNs, from the perspective of graph signal filtering. Tensor representation of lattice-structured graphs is next considered, and it is shown that tensors (multidimensional data arrays) are a special class of graph signals, whereby the graph vertices reside on a high-dimensional regular lattice structure. This part of monograph concludes with two emerging applications in financial data processing and underground transportation networks modeling.

false

true

false

true

false

159,210

2405.00903

A Named Entity Recognition and Topic Modeling-based Solution for Locating and Better Assessment of Natural Disasters in Social Media

Over the last decade, similar to other application domains, social media content has been proven very effective in disaster informatics. However, due to the unstructured nature of the data, several challenges are associated with disaster analysis in social media content. To fully explore the potential of social media content in disaster informatics, access to relevant content and the correct geo-location information is very critical. In this paper, we propose a three-step solution to tackling these challenges. Firstly, the proposed solution aims to classify social media posts into relevant and irrelevant posts followed by the automatic extraction of location information from the posts' text through Named Entity Recognition (NER) analysis. Finally, to quickly analyze the topics covered in large volumes of social media posts, we perform topic modeling resulting in a list of top keywords, that highlight the issues discussed in the tweet. For the Relevant Classification of Twitter Posts (RCTP), we proposed a merit-based fusion framework combining the capabilities of four different models namely BERT, RoBERTa, Distil BERT, and ALBERT obtaining the highest F1-score of 0.933 on a benchmark dataset. For the Location Extraction from Twitter Text (LETT), we evaluated four models namely BERT, RoBERTa, Distil BERTA, and Electra in an NER framework obtaining the highest F1-score of 0.960. For topic modeling, we used the BERTopic library to discover the hidden topic patterns in the relevant tweets. The experimental results of all the components of the proposed end-to-end solution are very encouraging and hint at the potential of social media content and NLP in disaster management.

false

true

false

451,122

2010.08285

Protograph-Based Low-Density Parity-Check Hadamard Codes

In this paper, we propose a new method to design low-density parity-check Hadamard (LDPC-Hadamard) codes, a type of ultimate-Shannon-limit approaching channel codes. The technique is based on applying Hadamard constraints to the check nodes in a generalized protograph-based LDPC code, followed by lifting the generalized protograph. We name the codes formed protograph-based LDPC Hadamard (PLDPC-Hadamard) codes. We also propose a modified Protograph Extrinsic Information Transfer (PEXIT) algorithm for analyzing and optimizing PLDPC-Hadamard code designs. The proposed algorithm further allows the analysis of PLDPC-Hadamard codes with degree-1 and/or punctured nodes. We find codes with decoding thresholds ranging from -1.53 dB to -1.42 dB. At a BER of 1e-5, the gaps of our codes to the ultimate-Shannon-limit range from 0.40 dB (for rate = 0.0494) to 0.16 dB (for rate = 0.003). Moreover, the error performance of our codes is comparable to that of the traditional LDPC-Hadamard codes. Finally, the BER performances of our codes after puncturing are simulated and compared.

false

true

false

201,139

2310.17316

Defect Spectrum: A Granular Look of Large-Scale Defect Datasets with Rich Semantics

Defect inspection is paramount within the closed-loop manufacturing system. However, existing datasets for defect inspection often lack precision and semantic granularity required for practical applications. In this paper, we introduce the Defect Spectrum, a comprehensive benchmark that offers precise, semantic-abundant, and large-scale annotations for a wide range of industrial defects. Building on four key industrial benchmarks, our dataset refines existing annotations and introduces rich semantic details, distinguishing multiple defect types within a single image. Furthermore, we introduce Defect-Gen, a two-stage diffusion-based generator designed to create high-quality and diverse defective images, even when working with limited datasets. The synthetic images generated by Defect-Gen significantly enhance the efficacy of defect inspection models. Overall, The Defect Spectrum dataset demonstrates its potential in defect inspection research, offering a solid platform for testing and refining advanced models.

false

true

false

403,082

1908.03298

On the Fundamental Limits of MIMO Massive Access Communication

The multiple access channel (MAC) with many-user is a general model for massive machine type communications. In this paradigm, the number of users may be comparable or even larger than the coding blocklength $n$. In contrast, classical MAC often assumes fixed and small number of the users. In this paper, we consider the massive access channel with multiple antennas system, where the base station (BS) with multiple receiving antennas serves the users in a single cell. The magnitude of users is assumed to grow unbounded with $n$. We investigate the achievable region of MIMO massive access channel, where among the total users, an unknown subset referred to active users may transmit data periodically. The asymptotic active user identification cost is also quantified. With the theoretical analysis, it was shown that given finite number of the receiving antennas, the individual rate for each user can be formulated as the sum rate multiplied by some specific factors, which correspond to the allocation of sum capacity. The successive decoding does not apply due to the interferences from growing unbounded users. Theoretical analysis shows that successive decoding works only when the number of receiving antennas goes to infinity with the increasing codelength.

false

true

false

141,214

1910.00193

Parallel Algorithm for Approximating Nash Equilibrium in Multiplayer Stochastic Games with Application to Naval Strategic Planning

Many real-world domains contain multiple agents behaving strategically with probabilistic transitions and uncertain (potentially infinite) duration. Such settings can be modeled as stochastic games. While algorithms have been developed for solving (i.e., computing a game-theoretic solution concept such as Nash equilibrium) two-player zero-sum stochastic games, research on algorithms for non-zero-sum and multiplayer stochastic games is limited. We present a new algorithm for these settings, which constitutes the first parallel algorithm for multiplayer stochastic games. We present experimental results on a 4-player stochastic game motivated by a naval strategic planning scenario, showing that our algorithm is able to quickly compute strategies constituting Nash equilibrium up to a very small degree of approximation error.

false

true

false

true

false

true

false

true

147,603

2112.02713

Joint Symmetry Detection and Shape Matching for Non-Rigid Point Cloud

Despite the success of deep functional maps in non-rigid 3D shape matching, there exists no learning framework that models both self-symmetry and shape matching simultaneously. This is despite the fact that errors due to symmetry mismatch are a major challenge in non-rigid shape matching. In this paper, we propose a novel framework that simultaneously learns both self symmetry as well as a pairwise map between a pair of shapes. Our key idea is to couple a self symmetry map and a pairwise map through a regularization term that provides a joint constraint on both of them, thereby, leading to more accurate maps. We validate our method on several benchmarks where it outperforms many competitive baselines on both tasks.

false

true

false

true

false

true

false

true

269,932

2402.09263

Uncertainty-Aware Transient Stability-Constrained Preventive Redispatch: A Distributional Reinforcement Learning Approach

Transient stability-constrained preventive redispatch plays a crucial role in ensuring power system security and stability. Since redispatch strategies need to simultaneously satisfy complex transient constraints and the economic need, model-based formulation and optimization become extremely challenging. In addition, the increasing uncertainty and variability introduced by renewable sources start to drive the system stability consideration from deterministic to probabilistic, which further exaggerates the complexity. In this paper, a Graph neural network guided Distributional Deep Reinforcement Learning (GD2RL) method is proposed, for the first time, to solve the uncertainty-aware transient stability-constrained preventive redispatch problem. First, a graph neural network-based transient simulator is trained by supervised learning to efficiently generate post-contingency rotor angle curves with the steady-state and contingency as inputs, which serves as a feature extractor for operating states and a surrogate time-domain simulator during the environment interaction for reinforcement learning. Distributional deep reinforcement learning with explicit uncertainty distribution of system operational conditions is then applied to generate the redispatch strategy to balance the user-specified probabilistic stability performance and economy preferences. The full distribution of the post-redispatch transient stability index is directly provided as the output. Case studies on the modified New England 39-bus system validate the proposed method.

false

true

false

429,440

1611.02256

A Big-Data Approach to Handle Many Process Variations: Tensor Recovery and Applications

Fabrication process variations are a major source of yield degradation in the nano-scale design of integrated circuits (IC), microelectromechanical systems (MEMS) and photonic circuits. Stochastic spectral methods are a promising technique to quantify the uncertainties caused by process variations. Despite their superior efficiency over Monte Carlo for many design cases, these algorithms suffer from the curse of dimensionality; i.e., their computational cost grows very fast as the number of random parameters increases. In order to solve this challenging problem, this paper presents a high-dimensional uncertainty quantification algorithm from a big-data perspective. Specifically, we show that the huge number of (e.g., $1.5 \times 10^{27}$) simulation samples in standard stochastic collocation can be reduced to a very small one (e.g., $500$) by exploiting some hidden structures of a high-dimensional data array. This idea is formulated as a tensor recovery problem with sparse and low-rank constraints; and it is solved with an alternating minimization approach. Numerical results show that our approach can simulate efficiently some ICs, as well as MEMS and photonic problems with over 50 independent random parameters, whereas the traditional algorithm can only handle several random parameters.

false

true

false

63,530

2409.07422

Controllable retinal image synthesis using conditional StyleGAN and latent space manipulation for improved diagnosis and grading of diabetic retinopathy

Diabetic retinopathy (DR) is a consequence of diabetes mellitus characterized by vascular damage within the retinal tissue. Timely detection is paramount to mitigate the risk of vision loss. However, training robust grading models is hindered by a shortage of annotated data, particularly for severe cases. This paper proposes a framework for controllably generating high-fidelity and diverse DR fundus images, thereby improving classifier performance in DR grading and detection. We achieve comprehensive control over DR severity and visual features (optic disc, vessel structure, lesion areas) within generated images solely through a conditional StyleGAN, eliminating the need for feature masks or auxiliary networks. Specifically, leveraging the SeFa algorithm to identify meaningful semantics within the latent space, we manipulate the DR images generated conditionally on grades, further enhancing the dataset diversity. Additionally, we propose a novel, effective SeFa-based data augmentation strategy, helping the classifier focus on discriminative regions while ignoring redundant features. Using this approach, a ResNet50 model trained for DR detection achieves 98.09% accuracy, 99.44% specificity, 99.45% precision, and an F1-score of 98.09%. Moreover, incorporating synthetic images generated by conditional StyleGAN into ResNet50 training for DR grading yields 83.33% accuracy, a quadratic kappa score of 87.64%, 95.67% specificity, and 72.24% precision. Extensive experiments conducted on the APTOS 2019 dataset demonstrate the exceptional realism of the generated images and the superior performance of our classifier compared to recent studies.

false

true

false

487,510

2210.05929

Few-shot Backdoor Attacks via Neural Tangent Kernels

In a backdoor attack, an attacker injects corrupted examples into the training set. The goal of the attacker is to cause the final trained model to predict the attacker's desired target label when a predefined trigger is added to test inputs. Central to these attacks is the trade-off between the success rate of the attack and the number of corrupted training examples injected. We pose this attack as a novel bilevel optimization problem: construct strong poison examples that maximize the attack success rate of the trained model. We use neural tangent kernels to approximate the training dynamics of the model being attacked and automatically learn strong poison examples. We experiment on subclasses of CIFAR-10 and ImageNet with WideResNet-34 and ConvNeXt architectures on periodic and patch trigger attacks and show that NTBA-designed poisoned examples achieve, for example, an attack success rate of 90% with ten times smaller number of poison examples injected compared to the baseline. We provided an interpretation of the NTBA-designed attacks using the analysis of kernel linear regression. We further demonstrate a vulnerability in overparametrized deep neural networks, which is revealed by the shape of the neural tangent kernel.

false

true

false

true

false

323,064

2103.03571

Cycle Self-Training for Domain Adaptation

Mainstream approaches for unsupervised domain adaptation (UDA) learn domain-invariant representations to narrow the domain shift. Recently, self-training has been gaining momentum in UDA, which exploits unlabeled target data by training with target pseudo-labels. However, as corroborated in this work, under distributional shift in UDA, the pseudo-labels can be unreliable in terms of their large discrepancy from target ground truth. Thereby, we propose Cycle Self-Training (CST), a principled self-training algorithm that explicitly enforces pseudo-labels to generalize across domains. CST cycles between a forward step and a reverse step until convergence. In the forward step, CST generates target pseudo-labels with a source-trained classifier. In the reverse step, CST trains a target classifier using target pseudo-labels, and then updates the shared representations to make the target classifier perform well on the source data. We introduce the Tsallis entropy as a confidence-friendly regularization to improve the quality of target pseudo-labels. We analyze CST theoretically under realistic assumptions, and provide hard cases where CST recovers target ground truth, while both invariant feature learning and vanilla self-training fail. Empirical results indicate that CST significantly improves over the state-of-the-arts on visual recognition and sentiment analysis benchmarks.

false

true

false

true

false

223,332

1611.05003

Light Field Stitching for Extended Synthetic Aperture

Through capturing spatial and angular radiance distribution, light field cameras introduce new capabilities that are not possible with conventional cameras. So far in the light field imaging literature, the focus has been on the theory and applications of single light field capture. By combining multiple light fields, it is possible to obtain new capabilities and enhancements, and even exceed physical limitations, such as spatial resolution and aperture size of the imaging device. In this paper, we present an algorithm to register and stitch multiple light fields. We utilize the regularity of the spatial and angular sampling in light field data, and extend some techniques developed for stereo vision systems to light field data. Such an extension is not straightforward for a micro-lens array (MLA) based light field camera due to extremely small baseline and low spatial resolution. By merging multiple light fields captured by an MLA based camera, we obtain larger synthetic aperture, which results in improvements in light field capabilities, such as increased depth estimation range/accuracy and wider perspective shift range.

false

true

false

63,936

1805.01891

Power Law in Sparsified Deep Neural Networks

The power law has been observed in the degree distributions of many biological neural networks. Sparse deep neural networks, which learn an economical representation from the data, resemble biological neural networks in many ways. In this paper, we study if these artificial networks also exhibit properties of the power law. Experimental results on two popular deep learning models, namely, multilayer perceptrons and convolutional neural networks, are affirmative. The power law is also naturally related to preferential attachment. To study the dynamical properties of deep networks in continual learning, we propose an internal preferential attachment model to explain how the network topology evolves. Experimental results show that with the arrival of a new task, the new connections made follow this preferential attachment process.

false

true

false

96,722

2312.17199

Tractable Function-Space Variational Inference in Bayesian Neural Networks

Reliable predictive uncertainty estimation plays an important role in enabling the deployment of neural networks to safety-critical settings. A popular approach for estimating the predictive uncertainty of neural networks is to define a prior distribution over the network parameters, infer an approximate posterior distribution, and use it to make stochastic predictions. However, explicit inference over neural network parameters makes it difficult to incorporate meaningful prior information about the data-generating process into the model. In this paper, we pursue an alternative approach. Recognizing that the primary object of interest in most settings is the distribution over functions induced by the posterior distribution over neural network parameters, we frame Bayesian inference in neural networks explicitly as inferring a posterior distribution over functions and propose a scalable function-space variational inference method that allows incorporating prior information and results in reliable predictive uncertainty estimates. We show that the proposed method leads to state-of-the-art uncertainty estimation and predictive performance on a range of prediction tasks and demonstrate that it performs well on a challenging safety-critical medical diagnosis task in which reliable uncertainty estimation is essential.

false

true

false

true

false

418,647

1007.3884

New Results for the MAP Problem in Bayesian Networks

This paper presents new results for the (partial) maximum a posteriori (MAP) problem in Bayesian networks, which is the problem of querying the most probable state configuration of some of the network variables given evidence. First, it is demonstrated that the problem remains hard even in networks with very simple topology, such as binary polytrees and simple trees (including the Naive Bayes structure). Such proofs extend previous complexity results for the problem. Inapproximability results are also derived in the case of trees if the number of states per variable is not bounded. Although the problem is shown to be hard and inapproximable even in very simple scenarios, a new exact algorithm is described that is empirically fast in networks of bounded treewidth and bounded number of states per variable. The same algorithm is used as basis of a Fully Polynomial Time Approximation Scheme for MAP under such assumptions. Approximation schemes were generally thought to be impossible for this problem, but we show otherwise for classes of networks that are important in practice. The algorithms are extensively tested using some well-known networks as well as random generated cases to show their effectiveness.

false

true

false

true

7,099

cs/0608093

Connection between continuous and digital n-manifolds and the Poincare conjecture

We introduce LCL covers of closed n-dimensional manifolds by n-dimensional disks and study their properties. We show that any LCL cover of an n-dimensional sphere can be converted to the minimal LCL cover, which consists of 2n+2 disks. We prove that an LCL collection of n-disks is a cover of a continuous n-sphere if and only if the intersection graph of this collection is a digital n-sphere. Using a link between LCL covers of closed continuous n-manifolds and digital n-manifolds, we find conditions where a continuous closed three-dimensional manifold is the three-dimensional sphere. We discuss a connection between the classification problems for closed continuous three-dimensional manifolds and digital three-manifolds.

false

true

false

true

539,655

2411.10513

Any2Any: Incomplete Multimodal Retrieval with Conformal Prediction

Autonomous agents perceive and interpret their surroundings by integrating multimodal inputs, such as vision, audio, and LiDAR. These perceptual modalities support retrieval tasks, such as place recognition in robotics. However, current multimodal retrieval systems encounter difficulties when parts of the data are missing due to sensor failures or inaccessibility, such as silent videos or LiDAR scans lacking RGB information. We propose Any2Any-a novel retrieval framework that addresses scenarios where both query and reference instances have incomplete modalities. Unlike previous methods limited to the imputation of two modalities, Any2Any handles any number of modalities without training generative models. It calculates pairwise similarities with cross-modal encoders and employs a two-stage calibration process with conformal prediction to align the similarities. Any2Any enables effective retrieval across multimodal datasets, e.g., text-LiDAR and text-time series. It achieves a Recall@5 of 35% on the KITTI dataset, which is on par with baseline models with complete modalities.

false

true

false

true

false

true

508,679

2309.01983

Quaternary Conjucyclic Codes with an Application to EAQEC Codes

Conjucyclic codes are part of a family of codes that includes cyclic, constacyclic, and quasi-cyclic codes, among others. Despite their importance in quantum error correction, they have not received much attention in the literature. This paper focuses on additive conjucyclic (ACC) codes over $\mathbb{F}_4$ and investigates their properties. Specifically, we derive the duals of ACC codes using a trace inner product and obtain the trace hull and its dimension. Also, establish a necessary and sufficient condition for an additive code to have a complementary dual (ACD). Additionally, we identify a necessary condition for an additive conjucyclic complementary pair of codes over $\mathbb{F}_4$. Furthermore, we show that the trace code of an ACC code is cyclic and provide a condition for the trace code of an ACC code to be LCD. To demonstrate the practical application of our findings, we construct some good entanglement-assisted quantum error-correcting (EAQEC) codes using the trace code of ACC codes.

false

true

false

389,878

2102.13185

Off-Policy Imitation Learning from Observations

Learning from Observations (LfO) is a practical reinforcement learning scenario from which many applications can benefit through the reuse of incomplete resources. Compared to conventional imitation learning (IL), LfO is more challenging because of the lack of expert action guidance. In both conventional IL and LfO, distribution matching is at the heart of their foundation. Traditional distribution matching approaches are sample-costly which depend on on-policy transitions for policy learning. Towards sample-efficiency, some off-policy solutions have been proposed, which, however, either lack comprehensive theoretical justifications or depend on the guidance of expert actions. In this work, we propose a sample-efficient LfO approach that enables off-policy optimization in a principled manner. To further accelerate the learning procedure, we regulate the policy update with an inverse action model, which assists distribution matching from the perspective of mode-covering. Extensive empirical results on challenging locomotion tasks indicate that our approach is comparable with state-of-the-art in terms of both sample-efficiency and asymptotic performance.

false

true

false

true

false

221,972

2312.03987

Cost-Effective In-Context Learning for Entity Resolution: A Design Space Exploration

Entity resolution (ER) is an important data integration task with a wide spectrum of applications. The state-of-the-art solutions on ER rely on pre-trained language models (PLMs), which require fine-tuning on a lot of labeled matching/non-matching entity pairs. Recently, large languages models (LLMs), such as GPT-4, have shown the ability to perform many tasks without tuning model parameters, which is known as in-context learning (ICL) that facilitates effective learning from a few labeled input context demonstrations. However, existing ICL approaches to ER typically necessitate providing a task description and a set of demonstrations for each entity pair and thus have limitations on the monetary cost of interfacing LLMs. To address the problem, in this paper, we provide a comprehensive study to investigate how to develop a cost-effective batch prompting approach to ER. We introduce a framework BATCHER consisting of demonstration selection and question batching and explore different design choices that support batch prompting for ER. We also devise a covering-based demonstration selection strategy that achieves an effective balance between matching accuracy and monetary cost. We conduct a thorough evaluation to explore the design space and evaluate our proposed strategies. Through extensive experiments, we find that batch prompting is very cost-effective for ER, compared with not only PLM-based methods fine-tuned with extensive labeled data but also LLM-based methods with manually designed prompting. We also provide guidance for selecting appropriate design choices for batch prompting.

false

true

false

true

false

413,497

2102.07638

AI Uncertainty Based on Rademacher Complexity and Shannon Entropy

In this paper from communication channel coding perspective we are able to present both a theoretical and practical discussion of AI's uncertainty, capacity and evolution for pattern classification based on the classical Rademacher complexity and Shannon entropy. First AI capacity is defined as in communication channels. It is shown qualitatively that the classical Rademacher complexity and Shannon entropy used in communication theory is closely related by their definitions, given a pattern classification problem with a complexity measured by Rademacher complexity. Secondly based on the Shannon mathematical theory on communication coding, we derive several sufficient and necessary conditions for an AI's error rate approaching zero in classifications problems. A 1/2 criteria on Shannon entropy is derived in this paper so that error rate can approach zero or is zero for AI pattern classification problems. Last but not least, we show our analysis and theory by providing examples of AI pattern classifications with error rate approaching zero or being zero.

false

true

false

true

false

220,169

2206.11309

GODEL: Large-Scale Pre-Training for Goal-Directed Dialog

We introduce GODEL (Grounded Open Dialogue Language Model), a large pre-trained language model for dialog. In contrast with earlier models such as DialoGPT, GODEL leverages a new phase of grounded pre-training designed to better support adapting GODEL to a wide range of downstream dialog tasks that require information external to the current conversation (e.g., a database or document) to produce good responses. Experiments against an array of benchmarks that encompass task-oriented dialog, conversational QA, and grounded open-domain dialog show that GODEL outperforms state-of-the-art pre-trained dialog models in few-shot fine-tuning setups, in terms of both human and automatic evaluation. A novel feature of our evaluation methodology is the introduction of a notion of utility that assesses the usefulness of responses (extrinsic evaluation) in addition to their communicative features (intrinsic evaluation). We show that extrinsic evaluation offers improved inter-annotator agreement and correlation with automated metrics. Code and data processing scripts are publicly available.

false

true

false

304,224

1907.01154

Adaptive Music Composition for Games

The generation of music that adapts dynamically to content and actions has an important role in building more immersive, memorable and emotive game experiences. To date, the development of adaptive music systems for video games is limited by both the nature of algorithms used for real-time music generation and the limited modelling of player action, game world context and emotion in current games. We propose that these issues must be addressed in tandem for the quality and flexibility of adaptive game music to significantly improve. Cognitive models of knowledge organisation and emotional affect are integrated with multi-modal, multi-agent composition techniques to produce a novel Adaptive Music System (AMS). The system is integrated into two stylistically distinct games. Gamers reported an overall higher immersion and correlation of music with game-world concepts with the AMS than with the original game soundtracks in both games.

false

true

false

true

false

true

137,244

0905.0374

Interference Alignment with Limited Feedback

We consider single-antenna interference networks where M sources, each with an average transmit power of P/M, communicate with M destinations over frequency-selective channels (with L taps each) and each destination has perfect knowledge of its channels from each of the sources. Assuming that there exist error-free non-interfering broadcast feedback links from each destination to all the nodes (i.e., sources and destinations) in the network, we show that naive interference alignment, in conjunction with vector quantization of the impulse response coefficients according to the scheme proposed in Mukkavilli et al., IEEE Trans. IT, 2003, achieves full spatial multiplexing gain of M/2, provided that the number of feedback bits broadcast by each destination is at least M(L-1) log P.

false

true

false

3,631

2404.00466

Computation and Communication Efficient Lightweighting Vertical Federated Learning

The exploration of computational and communication efficiency within Federated Learning (FL) has emerged as a prominent and crucial field of study. While most existing efforts to enhance these efficiencies have focused on Horizontal FL, the distinct processes and model structures of Vertical FL preclude the direct application of Horizontal FL-based techniques. In response, we introduce the concept of Lightweight Vertical Federated Learning (LVFL), targeting both computational and communication efficiencies. This approach involves separate lightweighting strategies for the feature model, to improve computational efficiency, and for feature embedding, to enhance communication efficiency. Moreover, we establish a convergence bound for our LVFL algorithm, which accounts for both communication and computational lightweighting ratios. Our evaluation of the algorithm on a image classification dataset reveals that LVFL significantly alleviates computational and communication demands while preserving robust learning performance. This work effectively addresses the gaps in communication and computational efficiency within Vertical FL.

false

true

false

true

442,939

2009.06687

Methods of the Vehicle Re-identification

Most of researchers use the vehicle re-identification based on classification. This always requires an update with the new vehicle models in the market. In this paper, two types of vehicle re-identification will be presented. First, the standard method, which needs an image from the search vehicle. VRIC and VehicleID data set are suitable for training this module. It will be explained in detail how to improve the performance of this method using a trained network, which is designed for the classification. The second method takes as input a representative image of the search vehicle with similar make/model, released year and colour. It is very useful when an image from the search vehicle is not available. It produces as output a shape and a colour features. This could be used by the matching across a database to re-identify vehicles, which look similar to the search vehicle. To get a robust module for the re-identification, a fine-grained classification has been trained, which its class consists of four elements: the make of a vehicle refers to the vehicle's manufacturer, e.g. Mercedes-Benz, the model of a vehicle refers to type of model within that manufacturer's portfolio, e.g. C Class, the year refers to the iteration of the model, which may receive progressive alterations and upgrades by its manufacturer and the perspective of the vehicle. Thus, all four elements describe the vehicle at increasing degree of specificity. The aim of the vehicle shape classification is to classify the combination of these four elements. The colour classification has been separately trained. The results of vehicle re-identification will be shown. Using a developed tool, the re-identification of vehicles on video images and on controlled data set will be demonstrated. This work was partially funded under the grant.

false

true

false

195,713

0902.3196

Symbolic Computing with Incremental Mindmaps to Manage and Mine Data Streams - Some Applications

In our understanding, a mind-map is an adaptive engine that basically works incrementally on the fundament of existing transactional streams. Generally, mind-maps consist of symbolic cells that are connected with each other and that become either stronger or weaker depending on the transactional stream. Based on the underlying biologic principle, these symbolic cells and their connections as well may adaptively survive or die, forming different cell agglomerates of arbitrary size. In this work, we intend to prove mind-maps' eligibility following diverse application scenarios, for example being an underlying management system to represent normal and abnormal traffic behaviour in computer networks, supporting the detection of the user behaviour within search engines, or being a hidden communication layer for natural language interaction.

false

true

false

true

false

3,191

2002.05308

Efficient Adaptive Experimental Design for Average Treatment Effect Estimation

We study how to efficiently estimate average treatment effects (ATEs) using adaptive experiments. In adaptive experiments, experimenters sequentially assign treatments to experimental units while updating treatment assignment probabilities based on past data. We start by defining the efficient treatment-assignment probability, which minimizes the semiparametric efficiency bound for ATE estimation. Our proposed experimental design estimates and uses the efficient treatment-assignment probability to assign treatments. At the end of the proposed design, the experimenter estimates the ATE using a newly proposed Adaptive Augmented Inverse Probability Weighting (A2IPW) estimator. We show that the asymptotic variance of the A2IPW estimator using data from the proposed design achieves the minimized semiparametric efficiency bound. We also analyze the estimator's finite-sample properties and develop nonparametric and nonasymptotic confidence intervals that are valid at any round of the proposed design. These anytime valid confidence intervals allow us to conduct rate-optimal sequential hypothesis testing, allowing for early stopping and reducing necessary sample size.

false

true

false

163,856

2401.14534

Meta-Learning Linear Quadratic Regulators: A Policy Gradient MAML Approach for Model-free LQR

We investigate the problem of learning linear quadratic regulators (LQR) in a multi-task, heterogeneous, and model-free setting. We characterize the stability and personalization guarantees of a policy gradient-based (PG) model-agnostic meta-learning (MAML) (Finn et al., 2017) approach for the LQR problem under different task-heterogeneity settings. We show that our MAML-LQR algorithm produces a stabilizing controller close to each task-specific optimal controller up to a task-heterogeneity bias in both model-based and model-free learning scenarios. Moreover, in the model-based setting, we show that such a controller is achieved with a linear convergence rate, which improves upon sub-linear rates from existing work. Our theoretical guarantees demonstrate that the learned controller can efficiently adapt to unseen LQR tasks.

false

true

false

424,127

1603.05355

GeoReach: An Efficient Approach for Evaluating Graph Reachability Queries with Spatial Range Predicates

Graphs are widely used to model data in many application domains. Thanks to the wide spread use of GPS-enabled devices, many applications assign a spatial attribute to graph vertices (e.g., geo-tagged social media). Users may issue a Reachability Query with Spatial Range Predicate (abbr. RangeReach). RangeReach finds whether an input vertex can reach any spatial vertex that lies within an input spatial range. An example of a RangeReach query is: Given a social graph, find whether Alice can reach any of the venues located within the geographical area of Arizona State University. The paper proposes GeoReach an approach that adds spatial data awareness to a graph database management system (GDBMS). GeoReach allows efficient execution of RangeReach queries, yet without compromising a lot on the overall system scalability (measured in terms of storage size and initialization/maintenance time). To achieve that, GeoReach is equipped with a light-weight data structure, namely SPA-Graph, that augments the underlying graph data with spatial indexing directories. When a RangeReach query is issued, the system employs a pruned-graph traversal approach. Experiments based on real system implementation inside Neo4j proves that GEOREACH exhibits up to two orders of magnitude better query response time and up to four times less storage than the state-of-the-art spatial and reachability indexing approaches.

false

true

false

true

false

53,352

2307.07691

A Survey on Change Detection Techniques in Document Images

The problem of change detection in images finds application in different domains like diagnosis of diseases in the medical field, detecting growth patterns of cities through remote sensing, and finding changes in legal documents and contracts. However, this paper presents a survey on core techniques and rules to detect changes in different versions of a document image. Our discussions on change detection focus on two categories -- content-based and layout-based. The content-based techniques intelligently extract and analyze the image contents (text or non-text) to show the possible differences, whereas the layout-based techniques use structural information to predict document changes. We also summarize the existing datasets and evaluation metrics used in change detection experiments. The shortcomings and challenges the existing methods face are reported, along with some pointers for future research work.

false

true

false

true

false

379,504

2404.14007

Infusion: Preventing Customized Text-to-Image Diffusion from Overfitting

Text-to-image (T2I) customization aims to create images that embody specific visual concepts delineated in textual descriptions. However, existing works still face a main challenge, concept overfitting. To tackle this challenge, we first analyze overfitting, categorizing it into concept-agnostic overfitting, which undermines non-customized concept knowledge, and concept-specific overfitting, which is confined to customize on limited modalities, i.e, backgrounds, layouts, styles. To evaluate the overfitting degree, we further introduce two metrics, i.e, Latent Fisher divergence and Wasserstein metric to measure the distribution changes of non-customized and customized concept respectively. Drawing from the analysis, we propose Infusion, a T2I customization method that enables the learning of target concepts to avoid being constrained by limited training modalities, while preserving non-customized knowledge. Remarkably, Infusion achieves this feat with remarkable efficiency, requiring a mere 11KB of trained parameters. Extensive experiments also demonstrate that our approach outperforms state-of-the-art methods in both single and multi-concept customized generation.

false

true

false

true

false

448,532

2109.10243

Beam Refinement and User State Acquisition via Integrated Sensing and Communication with OFDM

The performance of millimeter wave (mmWave) communications strongly relies on accurate beamforming both at base station and user terminal sides, referred to as beam alignment (BA). Existing BA algorithms provide initial yet coarse angle estimates as they typically use a codebook of a finite number of discreteized beams (angles). Towards emerging applications requiring timely and precise tracking of dynamically changing state of users, we consider a system where a base station with a co-located radar receiver estimates relevant state parameters of users and simultaneously sends OFDM-modulated data symbols. In particular, based on a hybrid digital analog data transmitter/radar receiver architecture, we propose a simple beam refinement and initial state acquisition scheme that can be used for beam and user location tracking in a dynamic environment. Numerical results inspired by IEEE802.11ad parameters demonstrate that the proposed method is able to improve significantly the communication rate and further achieve accurate state estimation.

false

true

false

256,553

2204.04059

Deep Learning-Based Intra Mode Derivation for Versatile Video Coding

In intra coding, Rate Distortion Optimization (RDO) is performed to achieve the optimal intra mode from a pre-defined candidate list. The optimal intra mode is also required to be encoded and transmitted to the decoder side besides the residual signal, where lots of coding bits are consumed. To further improve the performance of intra coding in Versatile Video Coding (VVC), an intelligent intra mode derivation method is proposed in this paper, termed as Deep Learning based Intra Mode Derivation (DLIMD). In specific, the process of intra mode derivation is formulated as a multi-class classification task, which aims to skip the module of intra mode signaling for coding bits reduction. The architecture of DLIMD is developed to adapt to different quantization parameter settings and variable coding blocks including non-square ones, which are handled by one single trained model. Different from the existing deep learning based classification problems, the hand-crafted features are also fed into the intra mode derivation network besides the learned features from feature learning network. To compete with traditional method, one additional binary flag is utilized in the video codec to indicate the selected scheme with RDO. Extensive experimental results reveal that the proposed method can achieve 2.28%, 1.74%, and 2.18% bit rate reduction on average for Y, U, and V components on the platform of VVC test model, which outperforms the state-of-the-art works.

false

true

false

true

290,524

2402.02464

A Graph is Worth $K$ Words: Euclideanizing Graph using Pure Transformer

Can we model Non-Euclidean graphs as pure language or even Euclidean vectors while retaining their inherent information? The Non-Euclidean property have posed a long term challenge in graph modeling. Despite recent graph neural networks and graph transformers efforts encoding graphs as Euclidean vectors, recovering the original graph from vectors remains a challenge. In this paper, we introduce GraphsGPT, featuring an Graph2Seq encoder that transforms Non-Euclidean graphs into learnable Graph Words in the Euclidean space, along with a GraphGPT decoder that reconstructs the original graph from Graph Words to ensure information equivalence. We pretrain GraphsGPT on $100$M molecules and yield some interesting findings: (1) The pretrained Graph2Seq excels in graph representation learning, achieving state-of-the-art results on $8/9$ graph classification and regression tasks. (2) The pretrained GraphGPT serves as a strong graph generator, demonstrated by its strong ability to perform both few-shot and conditional graph generation. (3) Graph2Seq+GraphGPT enables effective graph mixup in the Euclidean space, overcoming previously known Non-Euclidean challenges. (4) The edge-centric pretraining framework GraphsGPT demonstrates its efficacy in graph domain tasks, excelling in both representation and generation. Code is available at \href{https://github.com/A4Bio/GraphsGPT}{GitHub}.

false

true

false

true

false

426,577

2103.08890

LabelGit: A Dataset for Software Repositories Classification using Attributed Dependency Graphs

Software repository hosting services contain large amounts of open-source software, with GitHub hosting more than 100 million repositories, from new to established ones. Given this vast amount of projects, there is a pressing need for a search based on the software's content and features. However, even though GitHub offers various solutions to aid software discovery, most repositories do not have any labels, reducing the utility of search and topic-based analysis. Moreover, classifying software modules is also getting more importance given the increase in Component-Based Software Development. However, previous work focused on software classification using keyword-based approaches or proxies for the project (e.g., README), which is not always available. In this work, we create a new annotated dataset of GitHub Java projects called LabelGit. Our dataset uses direct information from the source code, like the dependency graph and source code neural representations from the identifiers. Using this dataset, we hope to aid the development of solutions that do not rely on proxies but use the entire source code to perform classification.

false

true

false

true

225,019

2205.15173

Self-Supervised Pre-training of Vision Transformers for Dense Prediction Tasks

We present a new self-supervised pre-training of Vision Transformers for dense prediction tasks. It is based on a contrastive loss across views that compares pixel-level representations to global image representations. This strategy produces better local features suitable for dense prediction tasks as opposed to contrastive pre-training based on global image representation only. Furthermore, our approach does not suffer from a reduced batch size since the number of negative examples needed in the contrastive loss is in the order of the number of local features. We demonstrate the effectiveness of our pre-training strategy on two dense prediction tasks: semantic segmentation and monocular depth estimation.

false

true

false

299,631

1106.1325

Shearlets and Optimally Sparse Approximations

Multivariate functions are typically governed by anisotropic features such as edges in images or shock fronts in solutions of transport-dominated equations. One major goal both for the purpose of compression as well as for an efficient analysis is the provision of optimally sparse approximations of such functions. Recently, cartoon-like images were introduced in 2D and 3D as a suitable model class, and approximation properties were measured by considering the decay rate of the $L^2$ error of the best $N$-term approximation. Shearlet systems are to date the only representation system, which provide optimally sparse approximations of this model class in 2D as well as 3D. Even more, in contrast to all other directional representation systems, a theory for compactly supported shearlet frames was derived which moreover also satisfy this optimality benchmark. This chapter shall serve as an introduction to and a survey about sparse approximations of cartoon-like images by band-limited and also compactly supported shearlet frames as well as a reference for the state-of-the-art of this research field.

false

true

false

true

10,749

2204.06769

Learning topological defects formation with neural networks in a quantum phase transition

Neural networks possess formidable representational power, rendering them invaluable in solving complex quantum many-body systems. While they excel at analyzing static solutions, nonequilibrium processes, including critical dynamics during a quantum phase transition, pose a greater challenge for neural networks. To address this, we utilize neural networks and machine learning algorithms to investigate the time evolutions, universal statistics, and correlations of topological defects in a one-dimensional transverse-field quantum Ising model. Specifically, our analysis involves computing the energy of the system during a quantum phase transition following a linear quench of the transverse magnetic field strength. The excitation energies satisfy a power-law relation to the quench rate, indicating a proportional relationship between the excitation energy and the kink numbers. Moreover, we establish a universal power-law relationship between the first three cumulants of the kink numbers and the quench rate, indicating a binomial distribution of the kinks. Finally, the normalized kink-kink correlations are also investigated and it is found that the numerical values are consistent with the analytic formula.

false

true

false

291,444

2411.14647

Benchmarking Multimodal Models for Ukrainian Language Understanding Across Academic and Cultural Domains

While the evaluation of multimodal English-centric models is an active area of research with numerous benchmarks, there is a profound lack of benchmarks or evaluation suites for low- and mid-resource languages. We introduce ZNO-Vision, a comprehensive multimodal Ukrainian-centric benchmark derived from standardized university entrance examination (ZNO). The benchmark consists of over 4,300 expert-crafted questions spanning 12 academic disciplines, including mathematics, physics, chemistry, and humanities. We evaluated the performance of both open-source models and API providers, finding that only a handful of models performed above baseline. Alongside the new benchmark, we performed the first evaluation study of multimodal text generation for the Ukrainian language: we measured caption generation quality on the Multi30K-UK dataset, translated the VQA benchmark into Ukrainian, and measured performance degradation relative to original English versions. Lastly, we tested a few models from a cultural perspective on knowledge of national cuisine. We believe our work will advance multimodal generation capabilities for the Ukrainian language and our approach could be useful for other low-resource languages.

false

true

false

510,259

2203.04699

Gym-saturation: an OpenAI Gym environment for saturation provers

`gym-saturation` is an OpenAI Gym environment for reinforcement learning (RL) agents capable of proving theorems. Currently, only theorems written in a formal language of the Thousands of Problems for Theorem Provers (TPTP) library in clausal normal form (CNF) are supported. `gym-saturation` implements the 'given clause' algorithm (similar to the one used in Vampire and E Prover). Being written in Python, `gym-saturation` was inspired by PyRes. In contrast to the monolithic architecture of a typical Automated Theorem Prover (ATP), `gym-saturation` gives different agents opportunities to select clauses themselves and train from their experience. Combined with a particular agent, `gym-saturation` can work as an ATP. Even with a non trained agent based on heuristics, `gym-saturation` can find refutations for 688 (of 8257) CNF problems from TPTP v7.5.0.

false

true

false

284,564

1906.08464

A Hierarchical Architecture for Sequential Decision-Making in Autonomous Driving using Deep Reinforcement Learning

Tactical decision making is a critical feature for advanced driving systems, that incorporates several challenges such as complexity of the uncertain environment and reliability of the autonomous system. In this work, we develop a multi-modal architecture that includes the environmental modeling of ego surrounding and train a deep reinforcement learning (DRL) agent that yields consistent performance in stochastic highway driving scenarios. To this end, we feed the occupancy grid of the ego surrounding into the DRL agent and obtain the high-level sequential commands (i.e. lane change) to send them to lower-level controllers. We will show that dividing the autonomous driving problem into a multi-layer control architecture enables us to leverage the AI power to solve each layer separately and achieve an admissible reliability score. Comparing with end-to-end approaches, this architecture enables us to end up with a more reliable system which can be implemented in actual self-driving cars.

false

true

false

true

false

true

false

135,870

2310.19093

Extending the Cooperative Dual-Task Space in Conformal Geometric Algebra

In this work, we are presenting an extension of the cooperative dual-task space (CDTS) in conformal geometric algebra. The CDTS was first defined using dual quaternion algebra and is a well established framework for the simplified definition of tasks using two manipulators. By integrating conformal geometric algebra, we aim to further enhance the geometric expressiveness and thus simplify the modeling of various tasks. We show this formulation by first presenting the CDTS and then its extension that is based around a cooperative pointpair. This extension keeps all the benefits of the original formulation that is based on dual quaternions, but adds more tools for geometric modeling of the dual-arm tasks. We also present how this CGA-CDTS can be seamlessly integrated with an optimal control framework in geometric algebra that was derived in previous work. In the experiments, we demonstrate how to model different objectives and constraints using the CGA-CDTS. Using a setup of two Franka Emika robots we then show the effectiveness of our approach using model predictive control in real world experiments.

false

true

false

403,850

2406.08604

GRU-Net: Gaussian Attention Aided Dense Skip Connection Based MultiResUNet for Breast Histopathology Image Segmentation

Breast cancer is a major global health concern. Pathologists face challenges in analyzing complex features from pathological images, which is a time-consuming and labor-intensive task. Therefore, efficient computer-based diagnostic tools are needed for early detection and treatment planning. This paper presents a modified version of MultiResU-Net for histopathology image segmentation, which is selected as the backbone for its ability to analyze and segment complex features at multiple scales and ensure effective feature flow via skip connections. The modified version also utilizes the Gaussian distribution-based Attention Module (GdAM) to incorporate histopathology-relevant text information in a Gaussian distribution. The sampled features from the Gaussian text feature-guided distribution highlight specific spatial regions based on prior knowledge. Finally, using the Controlled Dense Residual Block (CDRB) on skip connections of MultiResU-Net, the information is transferred from the encoder layers to the decoder layers in a controlled manner using a scaling parameter derived from the extracted spatial features. We validate our approach on two diverse breast cancer histopathology image datasets: TNBC and MonuSeg, demonstrating superior segmentation performance compared to state-of-the-art methods. The code for our proposed model is available on https://github.com/AyushRoy2001/GRU-Net.

false

true

false

true

false

463,547

2410.20305

Accelerating Direct Preference Optimization with Prefix Sharing

Offline paired preference optimization algorithms have become a popular approach for fine-tuning on preference data, outperforming traditional supervised fine-tuning in various tasks. However, traditional implementations often involve redundant computations, especially for tasks with long shared prompts. We introduce prefix sharing for preference tuning, a novel technique that processes chosen and rejected responses as one sequence with a shared prefix. To prevent cross-response contamination, we use a custom block-sparse attention mask. Our method achieves $1.1$-$1.5\times$ improvement in training throughput on popular DPO datasets, without any effect on convergence. When combined with sequence packing, we observe consistent $1.3$-$1.6\times$ speedups, benefiting even datasets with smaller sequence lengths. While we focus on Direct Preference Optimization (DPO), our approach is applicable to other paired preference tuning methods. By enhancing computational efficiency, our work contributes to making preference-based fine-tuning more accessible for a wider range of applications and model sizes. We open-source our code at https://github.com/frankxwang/dpo-prefix-sharing.

false

true

false

true

false

502,754

2206.09426

ADBench: Anomaly Detection Benchmark

Given a long list of anomaly detection algorithms developed in the last few decades, how do they perform with regard to (i) varying levels of supervision, (ii) different types of anomalies, and (iii) noisy and corrupted data? In this work, we answer these key questions by conducting (to our best knowledge) the most comprehensive anomaly detection benchmark with 30 algorithms on 57 benchmark datasets, named ADBench. Our extensive experiments (98,436 in total) identify meaningful insights into the role of supervision and anomaly types, and unlock future directions for researchers in algorithm selection and design. With ADBench, researchers can easily conduct comprehensive and fair evaluations for newly proposed methods on the datasets (including our contributed ones from natural language and computer vision domains) against the existing baselines. To foster accessibility and reproducibility, we fully open-source ADBench and the corresponding results.

false

true

false

true

false

303,568

2110.10932

Subspace Detours Meet Gromov-Wasserstein

In the context of optimal transport methods, the subspace detour approach was recently presented by Muzellec and Cuturi (2019). It consists in building a nearly optimal transport plan in the measures space from an optimal transport plan in a wisely chosen subspace, onto which the original measures are projected. The contribution of this paper is to extend this category of methods to the Gromov-Wasserstein problem, which is a particular type of transport distance involving the inner geometry of the compared distributions. After deriving the associated formalism and properties, we also discuss a specific cost for which we can show connections with the Knothe-Rosenblatt rearrangement. We finally give an experimental illustration on a shape matching problem.

false

true

false

262,307

2404.09352

Counteracting Concept Drift by Learning with Future Malware Predictions

The accuracy of deployed malware-detection classifiers degrades over time due to changes in data distributions and increasing discrepancies between training and testing data. This phenomenon is known as the concept drift. While the concept drift can be caused by various reasons in general, new malicious files are created by malware authors with a clear intention of avoiding detection. The existence of the intention opens a possibility for predicting such future samples. Including predicted samples in training data should consequently increase the accuracy of the classifiers on new testing data. We compare two methods for predicting future samples: (1) adversarial training and (2) generative adversarial networks (GANs). The first method explicitly seeks for adversarial examples against the classifier that are then used as a part of training data. Similarly, GANs also generate synthetic training data. We use GANs to learn changes in data distributions within different time periods of training data and then apply these changes to generate samples that could be in testing data. We compare these prediction methods on two different datasets: (1) Ember public dataset and (2) the internal dataset of files incoming to Avast. We show that while adversarial training yields more robust classifiers, this method is not a good predictor of future malware in general. This is in contrast with previously reported positive results in different domains (including natural language processing and spam detection). On the other hand, we show that GANs can be successfully used as predictors of future malware. We specifically examine malware families that exhibit significant changes in their data distributions over time and the experimental results confirm that GAN-based predictions can significantly improve the accuracy of the classifier on new, previously unseen data.

false

true

false

true

false

446,636

1806.06387

Mind the gap: quantification of incomplete ablation patterns after pulmonary vein isolation using minimum path search

Pulmonary vein isolation (PVI) is a common procedure for the treatment of atrial fibrillation (AF). A successful isolation produces a continuous lesion (scar) completely encircling the veins that stops activation waves from propagating to the atrial body. Unfortunately, the encircling lesion is often incomplete, becoming a combination of scar and gaps of healthy tissue. These gaps are potential causes of AF recurrence, which requires a redo of the isolation procedure. Late-gadolinium enhanced cardiac magnetic resonance (LGE-CMR) is a non-invasive method that may also be used to detect gaps, but it is currently a time-consuming process, prone to high inter-observer variability. In this paper, we present a method to semi-automatically identify and quantify ablation gaps. Gap quantification is performed through minimum path search in a graph where every node is a scar patch and the edges are the geodesic distances between patches. We propose the Relative Gap Measure (RGM) to estimate the percentage of gap around a vein, which is defined as the ratio of the overall gap length and the total length of the path that encircles the vein. Additionally, an advanced version of the RGM has been developed to integrate gap quantification estimates from different scar segmentation techniques into a single figure-of-merit. Population-based statistical and regional analysis of gap distribution was performed using a standardised parcellation of the left atrium. We have evaluated our method on synthetic and clinical data from 50 AF patients who underwent PVI with radiofrequency ablation. The population-based analysis concluded that the left superior PV is more prone to lesion gaps while the left inferior PV tends to have less gaps (p<0.05 in both cases), in the processed data. This type of information can be very useful for the optimization and objective assessment of PVI interventions.

false

true

false

100,689

2007.08364

A high fidelity synthetic face framework for computer vision

Analysis of faces is one of the core applications of computer vision, with tasks ranging from landmark alignment, head pose estimation, expression recognition, and face recognition among others. However, building reliable methods requires time-consuming data collection and often even more time-consuming manual annotation, which can be unreliable. In our work we propose synthesizing such facial data, including ground truth annotations that would be almost impossible to acquire through manual annotation at the consistency and scale possible through use of synthetic data. We use a parametric face model together with hand crafted assets which enable us to generate training data with unprecedented quality and diversity (varying shape, texture, expression, pose, lighting, and hair).

false

true

false

true

false

187,604

2210.02033

Learning Video-independent Eye Contact Segmentation from In-the-Wild Videos

Human eye contact is a form of non-verbal communication and can have a great influence on social behavior. Since the location and size of the eye contact targets vary across different videos, learning a generic video-independent eye contact detector is still a challenging task. In this work, we address the task of one-way eye contact detection for videos in the wild. Our goal is to build a unified model that can identify when a person is looking at his gaze targets in an arbitrary input video. Considering that this requires time-series relative eye movement information, we propose to formulate the task as a temporal segmentation. Due to the scarcity of labeled training data, we further propose a gaze target discovery method to generate pseudo-labels for unlabeled videos, which allows us to train a generic eye contact segmentation model in an unsupervised way using in-the-wild videos. To evaluate our proposed approach, we manually annotated a test dataset consisting of 52 videos of human conversations. Experimental results show that our eye contact segmentation model outperforms the previous video-dependent eye contact detector and can achieve 71.88% framewise accuracy on our annotated test set. Our code and evaluation dataset are available at https://github.com/ut-vision/Video-Independent-ECS.

false

true

false

321,497

1609.00053

Analysis of the Self Projected Matching Pursuit Algorithm

The convergence and numerical analysis of a low memory implementation of the Orthogonal Matching Pursuit greedy strategy, which is termed Self Projected Matching Pursuit, is presented. This approach renders an iterative way of solving the least squares problem with much less storage requirement than direct linear algebra techniques. Hence, it appropriate for solving large linear systems. The analysis highlights its suitability within the class of well posed problems.

false

true

false

true

false

60,428

2409.12333

Scale-specific auxiliary multi-task contrastive learning for deep liver vessel segmentation

Extracting hepatic vessels from abdominal images is of high interest for clinicians since it allows to divide the liver into functionally-independent Couinaud segments. In this respect, an automated liver blood vessel extraction is widely summoned. Despite the significant growth in performance of semantic segmentation methodologies, preserving the complex multi-scale geometry of main vessels and ramifications remains a major challenge. This paper provides a new deep supervised approach for vessel segmentation, with a strong focus on representations arising from the different scales inherent to the vascular tree geometry. In particular, we propose a new clustering technique to decompose the tree into various scale levels, from tiny to large vessels. Then, we extend standard 3D UNet to multi-task learning by incorporating scale-specific auxiliary tasks and contrastive learning to encourage the discrimination between scales in the shared representation. Promising results, depicted in several evaluation metrics, are revealed on the public 3D-IRCADb dataset.

false

true

false

true

false

true

false

489,536

2102.13517

Detection of Alzheimer's Disease Using Graph-Regularized Convolutional Neural Network Based on Structural Similarity Learning of Brain Magnetic Resonance Images

Objective: This paper presents an Alzheimer's disease (AD) detection method based on learning structural similarity between Magnetic Resonance Images (MRIs) and representing this similarity as a graph. Methods: We construct the similarity graph using embedded features of the input image (i.e., Non-Demented (ND), Very Mild Demented (VMD), Mild Demented (MD), and Moderated Demented (MDTD)). We experiment and compare different dimension-reduction and clustering algorithms to construct the best similarity graph to capture the similarity between the same class images using the cosine distance as a similarity measure. We utilize the similarity graph to present (sample) the training data to a convolutional neural network (CNN). We use the similarity graph as a regularizer in the loss function of a CNN model to minimize the distance between the input images and their k-nearest neighbours in the similarity graph while minimizing the categorical cross-entropy loss between the training image predictions and the actual image class labels. Results: We conduct extensive experiments with several pre-trained CNN models and compare the results to other recent methods. Conclusion: Our method achieves superior performance on the testing dataset (accuracy = 0.986, area under receiver operating characteristics curve = 0.998, F1 measure = 0.987). Significance: The classification results show an improvement in the prediction accuracy compared to the other methods. We release all the code used in our experiments to encourage reproducible research in this area

false

true

false

true

false

true

false

222,084

2203.09476

Uncertainty with UAV Search of Multiple Goal-oriented Targets

This paper considers the complex problem of a team of UAVs searching targets under uncertainty. The goal of the UAV team is to find all of the moving targets as quickly as possible before they arrive at their selected goal. The uncertainty considered is threefold: First, the UAVs do not know the targets' locations and destinations. Second, the sensing capabilities of the UAVs are not perfect. Third, the targets' movement model is unknown. We suggest a real-time algorithmic framework for the UAVs, combining entropy and stochastic-temporal belief, that aims at optimizing the probability of a quick and successful detection of all of the targets. We have empirically evaluated the algorithmic framework, and have shown its efficiency and significant performance improvement compared to other solutions. Furthermore, we have evaluated our framework using Peer Designed Agents (PDAs), which are computer agents that simulate targets and show that our algorithmic framework outperforms other solutions in this scenario.

false

true

false

true

false

286,166

1911.08635

Robust Deep Neural Networks Inspired by Fuzzy Logic

Deep neural networks have achieved impressive performance and become the de-facto standard in many tasks. However, troubling phenomena such as adversarial and fooling examples suggest that the generalization they make is flawed. I argue that among the roots of the phenomena are two geometric properties of common deep learning architectures: their distributed nature and the connectedness of their decision regions. As a remedy, I propose new architectures inspired by fuzzy logic that combine several alternative design elements. Through experiments on MNIST and CIFAR-10, the new models are shown to be more local, better at rejecting noise samples, and more robust against adversarial examples. Ablation analyses reveal behaviors on adversarial examples that cannot be explained by the linearity hypothesis but are consistent with the hypothesis that logic-inspired traits create more robust models.

false

true

false

true

154,249

2410.19696

Age of Coded Updates In Gossip Networks Under Memory and Memoryless Schemes

We consider an information update system on a gossip network, where a source node encodes information into $n$ total keys such that any subset of at least $k+1$ keys can fully reconstruct the original information. This encoding process follows the principles of a $k$-out-of-$n$ threshold system. The encoded updates are then disseminated across the network through peer-to-peer communication. We have two different types of nodes in a network: subscriber nodes, which receive a unique key from the source node for every status update instantaneously, and nonsubscriber nodes, which receive a unique key for an update only if the node is selected by the source, and this selection is renewed for each update. For the message structure between nodes, we consider two different schemes: a memory scheme (in which the nodes keep the source's current and previous encrypted messages) and a memoryless scheme (in which the nodes are allowed to only keep the source's current message). We measure the timeliness of information updates by using a recent performance metric, the version age of information. We present explicit formulas for the time average AoI in a scalable homogeneous network as functions of the network parameters under a memoryless scheme. Additionally, we provide strict lower and upper bounds for the time average AoI under a memory scheme.

false

true

false

502,420

2011.00776

Incorporating Gas Pipeline Leakage Failure Modes in Risk Evaluation of Electricity-Gas Integrated Energy Systems

In the existing literatures for the risk evaluation of electricity-gas integrated energy system (EGIES), the impacts of gas leakage in pipelines are ignored. This paper presents a method to incorporate the failure modes of gas pipeline leakage in EGIES risk evaluation. A Markov state transition model of gas pipeline with multi-state and multi-mode transition process, and a bi-level Monte Carlo sampling method for this model are developed. A stochastic topology change based network model of EGIES considering the pipeline leakage failure modes is presented. The risk indices for EGIES based on the load shedding, including those specifically for gas leakage risks, are also proposed. An EGIES with a modified RBTS and a 7-node gas system was used to demonstrate an application of the proposed method and models. The results indicate that pipeline leakage failures have significant impacts on the risk of EGIES. Ignoring pipeline leakage failures in the risk evaluation of EGIES will result in an overly underestimation of system risk and most likely a misleading conclusion in system planning.

false

true

false

204,366

2105.03647

Informative and Representative Triplet Selection for Multilabel Remote Sensing Image Retrieval

Learning the similarity between remote sensing (RS) images forms the foundation for content-based RS image retrieval (CBIR). Recently, deep metric learning approaches that map the semantic similarity of images into an embedding (metric) space have been found very popular in RS. A common approach for learning the metric space relies on the selection of triplets of similar (positive) and dissimilar (negative) images to a reference image called as an anchor. Choosing triplets is a difficult task particularly for multi-label RS CBIR, where each training image is annotated by multiple class labels. To address this problem, in this paper we propose a novel triplet sampling method in the framework of deep neural networks (DNNs) defined for multi-label RS CBIR problems. The proposed method selects a small set of the most representative and informative triplets based on two main steps. In the first step, a set of anchors that are diverse to each other in the embedding space is selected from the current mini-batch using an iterative algorithm. In the second step, different sets of positive and negative images are chosen for each anchor by evaluating the relevancy, hardness and diversity of the images among each other based on a novel strategy. Experimental results obtained on two multi-label benchmark archives show that the selection of the most informative and representative triplets in the context of DNNs results in: i) reducing the computational complexity of the training phase of the DNNs without any significant loss on the performance; and ii) an increase in learning speed since informative triplets allow fast convergence. The code of the proposed method is publicly available at https://git.tu-berlin.de/rsim/image-retrieval-from-triplets.

false

true

false

234,213

2502.03729

Action-Free Reasoning for Policy Generalization

End-to-end imitation learning offers a promising approach for training robot policies. However, generalizing to new settings remains a significant challenge. Although large-scale robot demonstration datasets have shown potential for inducing generalization, they are resource-intensive to scale. In contrast, human video data is abundant and diverse, presenting an attractive alternative. Yet, these human-video datasets lack action labels, complicating their use in imitation learning. Existing methods attempt to extract grounded action representations (e.g., hand poses), but resulting policies struggle to bridge the embodiment gap between human and robot actions. We propose an alternative approach: leveraging language-based reasoning from human videos-essential for guiding robot actions-to train generalizable robot policies. Building on recent advances in reasoning-based policy architectures, we introduce Reasoning through Action-free Data (RAD). RAD learns from both robot demonstration data (with reasoning and action labels) and action-free human video data (with only reasoning labels). The robot data teaches the model to map reasoning to low-level actions, while the action-free data enhances reasoning capabilities. Additionally, we will release a new dataset of 3,377 human-hand demonstrations with reasoning annotations compatible with the Bridge V2 benchmark and aimed at facilitating future research on reasoning-driven robot learning. Our experiments show that RAD enables effective transfer across the embodiment gap, allowing robots to perform tasks seen only in action-free data. Furthermore, scaling up action-free reasoning data significantly improves policy performance and generalization to novel tasks. These results highlight the promise of reasoning-driven learning from action-free datasets for advancing generalizable robot control. Project page: https://rad-generalization.github.io

false

true

false

true

false

530,836

1610.09300

Globally Optimal Training of Generalized Polynomial Neural Networks with Nonlinear Spectral Methods

The optimization problem behind neural networks is highly non-convex. Training with stochastic gradient descent and variants requires careful parameter tuning and provides no guarantee to achieve the global optimum. In contrast we show under quite weak assumptions on the data that a particular class of feedforward neural networks can be trained globally optimal with a linear convergence rate with our nonlinear spectral method. Up to our knowledge this is the first practically feasible method which achieves such a guarantee. While the method can in principle be applied to deep networks, we restrict ourselves for simplicity in this paper to one and two hidden layer networks. Our experiments confirm that these models are rich enough to achieve good performance on a series of real-world datasets.

false

true

false

63,037

2008.12950

Path Planning Followed by Kinodynamic Smoothing for Multirotor Aerial Vehicles (MAVs)

We explore path planning followed by kinodynamic smoothing while ensuring the vehicle dynamics feasibility for MAVs. We have chosen a geometrically based motion planning technique \textquotedblleft RRT*\textquotedblright\; for this purpose. In the proposed technique, we modified original RRT* introducing an adaptive search space and a steering function which help to increase the consistency of the planner. Moreover, we propose multiple RRT* which generates a set of desired paths, provided that the optimal path is selected among them. Then, apply kinodynamic smoothing, which will result in dynamically feasible as well as obstacle-free path. Thereafter, a b spline-based trajectory is generated to maneuver vehicle autonomously in unknown environments. Finally, we have tested the proposed technique in various simulated environments.

false

true

false

true

false

193,718

2008.10516

Exoplanet Validation with Machine Learning: 50 new validated Kepler planets

Over 30% of the ~4000 known exoplanets to date have been discovered using 'validation', where the statistical likelihood of a transit arising from a false positive (FP), non-planetary scenario is calculated. For the large majority of these validated planets calculations were performed using the vespa algorithm (Morton et al. 2016). Regardless of the strengths and weaknesses of vespa, it is highly desirable for the catalogue of known planets not to be dependent on a single method. We demonstrate the use of machine learning algorithms, specifically a gaussian process classifier (GPC) reinforced by other models, to perform probabilistic planet validation incorporating prior probabilities for possible FP scenarios. The GPC can attain a mean log-loss per sample of 0.54 when separating confirmed planets from FPs in the Kepler threshold crossing event (TCE) catalogue. Our models can validate thousands of unseen candidates in seconds once applicable vetting metrics are calculated, and can be adapted to work with the active TESS mission, where the large number of observed targets necessitates the use of automated algorithms. We discuss the limitations and caveats of this methodology, and after accounting for possible failure modes newly validate 50 Kepler candidates as planets, sanity checking the validations by confirming them with vespa using up to date stellar information. Concerning discrepancies with vespa arise for many other candidates, which typically resolve in favour of our models. Given such issues, we caution against using single-method planet validation with either method until the discrepancies are fully understood.

false

true

false

193,021

1707.00569

An In-Depth Analysis of Visual Tracking with Siamese Neural Networks

This survey presents a deep analysis of the learning and inference capabilities in nine popular trackers. It is neither intended to study the whole literature nor is it an attempt to review all kinds of neural networks proposed for visual tracking. We focus instead on Siamese neural networks which are a promising starting point for studying the challenging problem of tracking. These networks integrate efficiently feature learning and the temporal matching and have so far shown state-of-the-art performance. In particular, the branches of Siamese networks, their layers connecting these branches, specific aspects of training and the embedding of these networks into the tracker are highlighted. Quantitative results from existing papers are compared with the conclusion that the current evaluation methodology shows problems with the reproducibility and the comparability of results. The paper proposes a novel Lisp-like formalism for a better comparison of trackers. This assumes a certain functional design and functional decomposition of trackers. The paper tries to give foundation for tracker design by a formulation of the problem based on the theory of machine learning and by the interpretation of a tracker as a decision function. The work concludes with promising lines of research and suggests future work.

false

true

false

76,373

2410.09576

The Future of Learning in the Age of Generative AI: Automated Question Generation and Assessment with Large Language Models

In recent years, large language models (LLMs) and generative AI have revolutionized natural language processing (NLP), offering unprecedented capabilities in education. This chapter explores the transformative potential of LLMs in automated question generation and answer assessment. It begins by examining the mechanisms behind LLMs, emphasizing their ability to comprehend and generate human-like text. The chapter then discusses methodologies for creating diverse, contextually relevant questions, enhancing learning through tailored, adaptive strategies. Key prompting techniques, such as zero-shot and chain-of-thought prompting, are evaluated for their effectiveness in generating high-quality questions, including open-ended and multiple-choice formats in various languages. Advanced NLP methods like fine-tuning and prompt-tuning are explored for their role in generating task-specific questions, despite associated costs. The chapter also covers the human evaluation of generated questions, highlighting quality variations across different methods and areas for improvement. Furthermore, it delves into automated answer assessment, demonstrating how LLMs can accurately evaluate responses, provide constructive feedback, and identify nuanced understanding or misconceptions. Examples illustrate both successful assessments and areas needing improvement. The discussion underscores the potential of LLMs to replace costly, time-consuming human assessments when appropriately guided, showcasing their advanced understanding and reasoning capabilities in streamlining educational processes.

false

true

false

true

false

497,660

2203.16640

Task-driven Modular Co-design of Vehicle Control Systems

When designing autonomous systems, we need to consider multiple trade-offs at various abstraction levels, and the choices of single (hardware and software) components need to be studied jointly. In this work we consider the problem of designing the control algorithm as well as the platform on which it is executed. In particular, we focus on vehicle control systems, and formalize state-of-the-art control schemes as monotone feasibility relations. We then show how, leveraging a monotone theory of co-design, we can study the embedding of control synthesis problems into the task-driven co-design problem of a robotic platform. The properties of the proposed approach are illustrated by considering urban driving scenarios. We show how, given a particular task, we can efficiently compute Pareto optimal design solutions.

false

true

false

true

false

288,859

2409.12517

Scaling FP8 training to trillion-token LLMs

We train, for the first time, large language models using FP8 precision on datasets up to 2 trillion tokens -- a 20-fold increase over previous limits. Through these extended training runs, we uncover critical instabilities in FP8 training that were not observable in earlier works with shorter durations. We trace these instabilities to outlier amplification by the SwiGLU activation function. Interestingly, we show, both analytically and empirically, that this amplification happens only over prolonged training periods, and link it to a SwiGLU weight alignment process. To address this newly identified issue, we introduce Smooth-SwiGLU, a novel modification that ensures stable FP8 training without altering function behavior. We also demonstrate, for the first time, FP8 quantization of both Adam optimizer moments. Combining these innovations, we successfully train a 7B parameter model using FP8 precision on 256 Intel Gaudi2 accelerators, achieving on-par results with the BF16 baseline while delivering up to a $\sim 34 \%$ throughput improvement. A reference implementation is supplied in https://github.com/Anonymous1252022/Megatron-DeepSpeed.

false

true

false

true

false

489,626

2304.09058

Revisiting k-NN for Fine-tuning Pre-trained Language Models

Pre-trained Language Models (PLMs), as parametric-based eager learners, have become the de-facto choice for current paradigms of Natural Language Processing (NLP). In contrast, k-Nearest-Neighbor (kNN) classifiers, as the lazy learning paradigm, tend to mitigate over-fitting and isolated noise. In this paper, we revisit kNN classifiers for augmenting the PLMs-based classifiers. From the methodological level, we propose to adopt kNN with textual representations of PLMs in two steps: (1) Utilize kNN as prior knowledge to calibrate the training process. (2) Linearly interpolate the probability distribution predicted by kNN with that of the PLMs' classifier. At the heart of our approach is the implementation of kNN-calibrated training, which treats predicted results as indicators for easy versus hard examples during the training process. From the perspective of the diversity of application scenarios, we conduct extensive experiments on fine-tuning, prompt-tuning paradigms and zero-shot, few-shot and fully-supervised settings, respectively, across eight diverse end-tasks. We hope our exploration will encourage the community to revisit the power of classical methods for efficient NLP. Code and datasets are available in https://github.com/zjunlp/Revisit-KNN.

false

true

false

true

false

true

false

358,918

2006.08700

A multi-stage looking-ahead holding strategy to stabilize a high-frequency bus line

If a bus line becomes unstable, passengers waiting time will be lengthened and buses capacities will be mismatched. To stabilize a high-frequency bus line, many holding strategies have been proposed. Among these strategies, some need to take oversimplified assumptions to simplify the formulation of a real bus line; some may choose a myopic holding time because they take into account only the local and currently available headway information. To overcome the above shortcomings, we proposed an adaptive holding strategy which continuously reduces the deviations of the instantaneous headways from the Dynamic Target Headway -DTH. By using DTH, the new strategy can make use of the global headway information to determine a proper holding time. To fully estimate the influence of a holding time on the operation of a bus line in a relative long time period, we introduced the multi-stage looking-ahead mechanism into our new strategy. A detailed bus line simulation system was used to realize the above mechanism and to describe all kinds of complex components of a bus line. The numerical experiment demonstrated the effectiveness of our new strategy. Some meaningful insights were uncovered as follow: a. The number of stages to be looked ahead should not be too small or too big; b. The bigger the range of the action set, the better the performance of our strategy; c. To obtain the optimal performance, the interval between two neighbor holding times should not be too small or too big; d. the bigger the number of the control points, the better the performance of our strategy.

false

true

false

182,268

1805.12017

Robustifying Models Against Adversarial Attacks by Langevin Dynamics

Adversarial attacks on deep learning models have compromised their performance considerably. As remedies, a lot of defense methods were proposed, which however, have been circumvented by newer attacking strategies. In the midst of this ensuing arms race, the problem of robustness against adversarial attacks still remains unsolved. This paper proposes a novel, simple yet effective defense strategy where adversarial samples are relaxed onto the underlying manifold of the (unknown) target class distribution. Specifically, our algorithm drives off-manifold adversarial samples towards high density regions of the data generating distribution of the target class by the Metroplis-adjusted Langevin algorithm (MALA) with perceptual boundary taken into account. Although the motivation is similar to projection methods, e.g., Defense-GAN, our algorithm, called MALA for DEfense (MALADE), is equipped with significant dispersion - projection is distributed broadly, and therefore any whitebox attack cannot accurately align the input so that the MALADE moves it to a targeted untrained spot where the model predicts a wrong label. In our experiments, MALADE exhibited state-of-the-art performance against various elaborate attacking strategies.

false

true

false

99,071

2211.15736

Post-training Quantization on Diffusion Models

Denoising diffusion (score-based) generative models have recently achieved significant accomplishments in generating realistic and diverse data. These approaches define a forward diffusion process for transforming data into noise and a backward denoising process for sampling data from noise. Unfortunately, the generation process of current denoising diffusion models is notoriously slow due to the lengthy iterative noise estimations, which rely on cumbersome neural networks. It prevents the diffusion models from being widely deployed, especially on edge devices. Previous works accelerate the generation process of diffusion model (DM) via finding shorter yet effective sampling trajectories. However, they overlook the cost of noise estimation with a heavy network in every iteration. In this work, we accelerate generation from the perspective of compressing the noise estimation network. Due to the difficulty of retraining DMs, we exclude mainstream training-aware compression paradigms and introduce post-training quantization (PTQ) into DM acceleration. However, the output distributions of noise estimation networks change with time-step, making previous PTQ methods fail in DMs since they are designed for single-time step scenarios. To devise a DM-specific PTQ method, we explore PTQ on DM in three aspects: quantized operations, calibration dataset, and calibration metric. We summarize and use several observations derived from all-inclusive investigations to formulate our method, which especially targets the unique multi-time-step structure of DMs. Experimentally, our method can directly quantize full-precision DMs into 8-bit models while maintaining or even improving their performance in a training-free manner. Importantly, our method can serve as a plug-and-play module on other fast-sampling methods, e.g., DDIM. The code is available at https://github.com/42Shawn/PTQ4DM .

false

true

false

333,373

1606.02825

Arbitrage-Free Combinatorial Market Making via Integer Programming

We present a new combinatorial market maker that operates arbitrage-free combinatorial prediction markets specified by integer programs. Although the problem of arbitrage-free pricing, while maintaining a bound on the subsidy provided by the market maker, is #P-hard in the worst case, we posit that the typical case might be amenable to modern integer programming (IP) solvers. At the crux of our method is the Frank-Wolfe (conditional gradient) algorithm which is used to implement a Bregman projection aligned with the market maker's cost function, using an IP solver as an oracle. We demonstrate the tractability and improved accuracy of our approach on real-world prediction market data from combinatorial bets placed on the 2010 NCAA Men's Division I Basketball Tournament, where the outcome space is of size 2^63. To our knowledge, this is the first implementation and empirical evaluation of an arbitrage-free combinatorial prediction market on this scale.

false

true

false

true

57,022

2403.04670

End-to-end Conditional Robust Optimization

The field of Contextual Optimization (CO) integrates machine learning and optimization to solve decision making problems under uncertainty. Recently, a risk sensitive variant of CO, known as Conditional Robust Optimization (CRO), combines uncertainty quantification with robust optimization in order to promote safety and reliability in high stake applications. Exploiting modern differentiable optimization methods, we propose a novel end-to-end approach to train a CRO model in a way that accounts for both the empirical risk of the prescribed decisions and the quality of conditional coverage of the contextual uncertainty set that supports them. While guarantees of success for the latter objective are impossible to obtain from the point of view of conformal prediction theory, high quality conditional coverage is achieved empirically by ingeniously employing a logistic regression differentiable layer within the calculation of coverage quality in our training loss. We show that the proposed training algorithms produce decisions that outperform the traditional estimate then optimize approaches.

false

true

false

435,684

2311.10764

Deep Group Interest Modeling of Full Lifelong User Behaviors for CTR Prediction

Extracting users' interests from their lifelong behavior sequence is crucial for predicting Click-Through Rate (CTR). Most current methods employ a two-stage process for efficiency: they first select historical behaviors related to the candidate item and then deduce the user's interest from this narrowed-down behavior sub-sequence. This two-stage paradigm, though effective, leads to information loss. Solely using users' lifelong click behaviors doesn't provide a complete picture of their interests, leading to suboptimal performance. In our research, we introduce the Deep Group Interest Network (DGIN), an end-to-end method to model the user's entire behavior history. This includes all post-registration actions, such as clicks, cart additions, purchases, and more, providing a nuanced user understanding. We start by grouping the full range of behaviors using a relevant key (like item_id) to enhance efficiency. This process reduces the behavior length significantly, from O(10^4) to O(10^2). To mitigate the potential loss of information due to grouping, we incorporate two categories of group attributes. Within each group, we calculate statistical information on various heterogeneous behaviors (like behavior counts) and employ self-attention mechanisms to highlight unique behavior characteristics (like behavior type). Based on this reorganized behavior data, the user's interests are derived using the Transformer technique. Additionally, we identify a subset of behaviors that share the same item_id with the candidate item from the lifelong behavior sequence. The insights from this subset reveal the user's decision-making process related to the candidate item, improving prediction accuracy. Our comprehensive evaluation, both on industrial and public datasets, validates DGIN's efficacy and efficiency.

false

true

false

408,638

2108.01548

Inference via Sparse Coding in a Hierarchical Vision Model

Sparse coding has been incorporated in models of the visual cortex for its computational advantages and connection to biology. But how the level of sparsity contributes to performance on visual tasks is not well understood. In this work, sparse coding has been integrated into an existing hierarchical V2 model (Hosoya and Hyv\"arinen, 2015), but replacing its independent component analysis (ICA) with an explicit sparse coding in which the degree of sparsity can be controlled. After training, the sparse coding basis functions with a higher degree of sparsity resembled qualitatively different structures, such as curves and corners. The contributions of the models were assessed with image classification tasks, specifically tasks associated with mid-level vision including figure-ground classification, texture classification, and angle prediction between two line stimuli. In addition, the models were assessed in comparison to a texture sensitivity measure that has been reported in V2 (Freeman et al., 2013), and a deleted-region inference task. The results from the experiments show that while sparse coding performed worse than ICA at classifying images, only sparse coding was able to better match the texture sensitivity level of V2 and infer deleted image regions, both by increasing the degree of sparsity in sparse coding. Higher degrees of sparsity allowed for inference over larger deleted image regions. The mechanism that allows for this inference capability in sparse coding is described here.

false

true

false

true

false

249,062

2410.06065

Posets and Bounded Probabilities for Discovering Order-inducing Features in Event Knowledge Graphs

Event knowledge graphs (EKG) extend the classical notion of a trace to capture multiple, interacting views of a process execution. In this paper, we tackle the open problem of automating EKG discovery from uncurated data through a principled, probabilistic framing based on the outcome space resulting from featured-derived partial orders on events. From this, we derive an EKG discovery algorithm based upon statistical inference rather than an ad-hoc or heuristic-based strategy, or relying on manual analysis from domain experts. This approach comes at the computational cost of exploring a large, non-convex hypothesis space. In particular, solving the maximum likelihood term involves counting the number of linear extensions of posets, which in general is #P-complete. Fortunately, bound estimates suffice for model comparison, and admit incorporation into a bespoke branch-and-bound algorithm. We show that the posterior probability as defined is antitonic w.r.t. search depth for branching rules that are monotonic w.r.t. model inclusion. This allows pruning of large portions of the search space, which we show experimentally leads to rapid convergence toward optimal solutions that are consistent with manually built EKGs.

false

true

false

true

false

496,037

2410.15425

Accelerated Sub-Image Search For Variable-Size Patches Identification Based On Virtual Time Series Transformation And Segmentation

This paper addresses two tasks: (i) fixed-size objects such as hay bales are to be identified in an aerial image for a given reference image of the object, and (ii) variable-size patches such as areas on fields requiring spot spraying or other handling are to be identified in an image for a given small-scale reference image. Both tasks are related. The second differs in that identified sub-images similar to the reference image are further clustered before patches contours are determined by solving a traveling salesman problem. Both tasks are complex in that the exact number of similar sub-images is not known a priori. The main discussion of this paper is presentation of an acceleration mechanism for sub-image search that is based on a transformation of an image to multivariate time series along the RGB-channels and subsequent segmentation to reduce the 2D search space in the image. Two variations of the acceleration mechanism are compared to exhaustive search on diverse synthetic and real-world images. Quantitatively, proposed method results in solve time reductions of up to 2 orders of magnitude, while qualitatively delivering comparative visual results. Proposed method is neural network-free and does not use any image pre-processing.

false

true

false

true

false

500,531

2403.02922

From Spectra to Biophysical Insights: End-to-End Learning with a Biased Radiative Transfer Model

Advances in machine learning have boosted the use of Earth observation data for climate change research. Yet, the interpretability of machine-learned representations remains a challenge, particularly in understanding forests' biophysical reactions to climate change. Traditional methods in remote sensing that invert radiative transfer models (RTMs) to retrieve biophysical variables from spectral data often fail to account for biases inherent in the RTM, especially for complex forests. We propose to integrate RTMs into an auto-encoder architecture, creating an end-to-end learning approach. Our method not only corrects biases in RTMs but also outperforms traditional techniques for variable retrieval like neural network regression. Furthermore, our framework has potential generally for inverting biased physical models. The code is available on https://github.com/yihshe/ai-refined-rtm.git.

false

true

false

434,995

2402.04275

Motion Mapping Cognition: A Nondecomposable Primary Process in Human Vision

Human intelligence seems so mysterious that we have not successfully understood its foundation until now. Here, I want to present a basic cognitive process, motion mapping cognition (MMC), which should be a nondecomposable primary function in human vision. Wherein, I point out that, MMC process can be used to explain most of human visual functions in fundamental, but can not be effectively modelled by traditional visual processing ways including image segmentation, object recognition, object tracking etc. Furthermore, I state that MMC may be looked as an extension of Chen's theory of topological perception on human vision, and seems to be unsolvable using existing intelligent algorithm skills. Finally, along with the requirements of MMC problem, an interesting computational model, quantized topological matching principle can be derived by developing the idea of optimal transport theory. Above results may give us huge inspiration to develop more robust and interpretable machine vision models.

false

true

false

true

false

427,387

1303.7034

Energy Efficient Cooperative Strategies for Relay-Assisted Downlink Cellular Systems Part II: Practical Design

In a companion paper [1], we present a general approach to evaluate the impact of cognition in a downlink cellular system in which multiple relays assist the transmission of the base station. This approach is based on a novel theoretical tool which produces transmission schemes involving rate-splitting, superposition coding and interference decoding for a network with any number of relays and receivers. This second part focuses on a practical design example for a network in which a base station transmits to three receivers with the aid of two relay nodes. For this simple network, we explicitly evaluate the impact of relay cognition and precisely characterize the trade offs between the total energy consumption and the rate improvements provided by relay cooperation. These closedform expressions provide important insights on the role of cognition in larger networks and highlights interesting interference management strategies. We also present a numerical simulation setup in which we fully automate the derivation of achievable rate region for a general relay-assisted downlink cellular network. Our simulations clearly show the great advantages provided by cooperative strategies at the relays as compared to the uncoordinated scenario under varying channel conditions and target rates. These results are obtained by considering a large number of transmission strategies for different levels of relay cognition and numerically determining one that is the most energy efficient. The limited computational complexity of the numerical evaluations makes this approach suitable for the optimization of transmission strategies for larger networks.

false

true

false

23,312

2409.20530

Dual Encoder GAN Inversion for High-Fidelity 3D Head Reconstruction from Single Images

3D GAN inversion aims to project a single image into the latent space of a 3D Generative Adversarial Network (GAN), thereby achieving 3D geometry reconstruction. While there exist encoders that achieve good results in 3D GAN inversion, they are predominantly built on EG3D, which specializes in synthesizing near-frontal views and is limiting in synthesizing comprehensive 3D scenes from diverse viewpoints. In contrast to existing approaches, we propose a novel framework built on PanoHead, which excels in synthesizing images from a 360-degree perspective. To achieve realistic 3D modeling of the input image, we introduce a dual encoder system tailored for high-fidelity reconstruction and realistic generation from different viewpoints. Accompanying this, we propose a stitching framework on the triplane domain to get the best predictions from both. To achieve seamless stitching, both encoders must output consistent results despite being specialized for different tasks. For this reason, we carefully train these encoders using specialized losses, including an adversarial loss based on our novel occlusion-aware triplane discriminator. Experiments reveal that our approach surpasses the existing encoder training methods qualitatively and quantitatively. Please visit the project page: https://berkegokmen1.github.io/dual-enc-3d-gan-inv.

false

true

false

true

false

true

493,158

1911.08963

Parallel Implementations for Computing the Minimum Distance of a Random Linear Code on Multicomputers

The minimum distance of a linear code is a key concept in information theory. Therefore, the time required by its computation is very important to many problems in this area. In this paper, we introduce a family of implementations of the Brouwer-Zimmermann algorithm for distributed-memory architectures for computing the minimum distance of a random linear code over F2. Both current commercial and public-domain software only work on either unicore architectures or shared-memory architectures, which are limited in the number of cores/processors employed in the computation. Our implementations focus on distributed-memory architectures, thus being able to employ hundreds or even thousands of cores in the computation of the minimum distance. Our experimental results show that our implementations are much faster, even up to several orders of magnitude, than current implementations widely used nowadays.

false

true

false

true

154,363

1805.05008

Integrating Hypertension Phenotype and Genotype with Hybrid Non-negative Matrix Factorization

Hypertension is a heterogeneous syndrome in need of improved subtyping using phenotypic and genetic measurements so that patients in different subtypes share similar pathophysiologic mechanisms and respond more uniformly to targeted treatments. Existing machine learning approaches often face challenges in integrating phenotype and genotype information and presenting to clinicians an interpretable model. We aim to provide informed patient stratification by introducing Hybrid Non-negative Matrix Factorization (HNMF) on phenotype and genotype matrices. HNMF simultaneously approximates the phenotypic and genetic matrices using different appropriate loss functions, and generates patient subtypes, phenotypic groups and genetic groups. Unlike previous methods, HNMF approximates phenotypic matrix under Frobenius loss, and genetic matrix under Kullback-Leibler (KL) loss. We propose an alternating projected gradient method to solve the approximation problem. Simulation shows HNMF converges fast and accurately to the true factor matrices. On real-world clinical dataset, we used the patient factor matrix as features to predict main cardiac mechanistic outcomes. We compared HNMF with six different models using phenotype or genotype features alone, with or without NMF, or using joint NMF with only one type of loss. HNMF significantly outperforms all comparison models. HNMF also reveals intuitive phenotype-genotype interactions that characterize cardiac abnormalities.

false

true

false

97,359

1907.01930

Interference Avoidance Position Planning in Dual-hop and Multi-hop UAV Relay Networks

We consider unmanned aerial vehicle (UAV)-assisted wireless communication employing UAVs as relay nodes to increase the throughput between a pair of transmitter and receiver. We focus on developing effective methods to position the UAV(s) in the sky in the presence of interference in the environment, the existence of which makes the problem non-trivial and our methodology different from the current art. We study the optimal position planning, which aims to maximize the (average) signal-to-interference-ratio (SIR) of the system, in the presence of: i) one major source of interference, ii) stochastic interference. For each scenario, we first consider utilizing a single UAV in the dual-hop relay mode and determine its optimal position. Afterward, multiple UAVs in the multi-hop relay mode are considered, for which we investigate two novel problems concerned with determining the optimal number of required UAVs and developing an optimal fully distributed position alignment method. Subsequently, we propose a cost-effective method that simultaneously minimizes the number of UAVs and determines their optimal positions to guarantee a certain (average) SIR of the system. Alternatively, for a given number of UAVs, we develop a fully distributed placement algorithm along with its performance guarantee. Numerical simulations are provided to evaluate the performance of our proposed methods.

false

true

false

true

137,478

1311.0320

An Improved Solution for Restricted and Uncertain TRQ

CSPTRQ is an interesting problem and its has attracted much attention. The CSPTRQ is a variant of the traditional PTRQ. As objects moving in a constrained-space are common, clearly, it can also find many applications. At the first sight, our problem can be easily tackled by extending existing methods used to answer the PTRQ. Unfortunately, those classical techniques are not well suitable for our problem, due to a set of new challenges. We develop targeted solutions and demonstrate the efficiency and effectiveness of the proposed methods through extensive experiments.

false

true

false

28,141

2103.10493

Image Synthesis for Data Augmentation in Medical CT using Deep Reinforcement Learning

Deep learning has shown great promise for CT image reconstruction, in particular to enable low dose imaging and integrated diagnostics. These merits, however, stand at great odds with the low availability of diverse image data which are needed to train these neural networks. We propose to overcome this bottleneck via a deep reinforcement learning (DRL) approach that is integrated with a style-transfer (ST) methodology, where the DRL generates the anatomical shapes and the ST synthesizes the texture detail. We show that our method bears high promise for generating novel and anatomically accurate high resolution CT images at large and diverse quantities. Our approach is specifically designed to work with even small image datasets which is desirable given the often low amount of image data many researchers have available to them.

false

true

false

225,469

2010.10915

Contrastive Learning of General-Purpose Audio Representations

We introduce COLA, a self-supervised pre-training approach for learning a general-purpose representation of audio. Our approach is based on contrastive learning: it learns a representation which assigns high similarity to audio segments extracted from the same recording while assigning lower similarity to segments from different recordings. We build on top of recent advances in contrastive learning for computer vision and reinforcement learning to design a lightweight, easy-to-implement self-supervised model of audio. We pre-train embeddings on the large-scale Audioset database and transfer these representations to 9 diverse classification tasks, including speech, music, animal sounds, and acoustic scenes. We show that despite its simplicity, our method significantly outperforms previous self-supervised systems. We furthermore conduct ablation studies to identify key design choices and release a library to pre-train and fine-tune COLA models.

false

true

false

true

false

202,056

1507.02045

What Your Username Says About You

Usernames are ubiquitous on the Internet, and they are often suggestive of user demographics. This work looks at the degree to which gender and language can be inferred from a username alone by making use of unsupervised morphology induction to decompose usernames into sub-units. Experimental results on the two tasks demonstrate the effectiveness of the proposed morphological features compared to a character n-gram baseline.

false

true

false

44,936

1912.01497

Robust and Secure Wireless Communications via Intelligent Reflecting Surfaces

In this paper, intelligent reflecting surfaces (IRSs) are employed to enhance the physical layer security in a challenging radio environment. In particular, a multi-antenna access point (AP) has to serve multiple single-antenna legitimate users, which do not have line-of-sight communication links, in the presence of multiple multi-antenna potential eavesdroppers whose channel state information (CSI) is not perfectly known. Artificial noise (AN) is transmitted from the AP to deliberately impair the eavesdropping channels for security provisioning. We investigate the joint design of the beamformers and AN covariance matrix at the AP and the phase shifters at the IRSs for maximization of the system sum-rate while limiting the maximum information leakage to the potential eavesdroppers. To this end, we formulate a robust nonconvex optimization problem taking into account the impact of the imperfect CSI of the eavesdropping channels. To address the non-convexity of the optimization problem, an efficient algorithm is developed by capitalizing on alternating optimization, a penalty-based approach, successive convex approximation, and semidefinite relaxation. Simulation results show that IRSs can significantly improve the system secrecy performance compared to conventional architectures without IRS. Furthermore, our results unveil that, for physical layer security, uniformly distributing the reflecting elements among multiple IRSs is preferable over deploying them at a single IRS.

false

true

false

156,104

2409.01266

Double Machine Learning meets Panel Data -- Promises, Pitfalls, and Potential Solutions

Estimating causal effect using machine learning (ML) algorithms can help to relax functional form assumptions if used within appropriate frameworks. However, most of these frameworks assume settings with cross-sectional data, whereas researchers often have access to panel data, which in traditional methods helps to deal with unobserved heterogeneity between units. In this paper, we explore how we can adapt double/debiased machine learning (DML) (Chernozhukov et al., 2018) for panel data in the presence of unobserved heterogeneity. This adaptation is challenging because DML's cross-fitting procedure assumes independent data and the unobserved heterogeneity is not necessarily additively separable in settings with nonlinear observed confounding. We assess the performance of several intuitively appealing estimators in a variety of simulations. While we find violations of the cross-fitting assumptions to be largely inconsequential for the accuracy of the effect estimates, many of the considered methods fail to adequately account for the presence of unobserved heterogeneity. However, we find that using predictive models based on the correlated random effects approach (Mundlak, 1978) within DML leads to accurate coefficient estimates across settings, given a sample size that is large relative to the number of observed confounders. We also show that the influence of the unobserved heterogeneity on the observed confounders plays a significant role for the performance of most alternative methods.

false

true

false

485,278

2404.00674

Knowledge NeRF: Few-shot Novel View Synthesis for Dynamic Articulated Objects

We present Knowledge NeRF to synthesize novel views for dynamic scenes. Reconstructing dynamic 3D scenes from few sparse views and rendering them from arbitrary perspectives is a challenging problem with applications in various domains. Previous dynamic NeRF methods learn the deformation of articulated objects from monocular videos. However, qualities of their reconstructed scenes are limited. To clearly reconstruct dynamic scenes, we propose a new framework by considering two frames at a time.We pretrain a NeRF model for an articulated object.When articulated objects moves, Knowledge NeRF learns to generate novel views at the new state by incorporating past knowledge in the pretrained NeRF model with minimal observations in the present state. We propose a projection module to adapt NeRF for dynamic scenes, learning the correspondence between pretrained knowledge base and current states. Experimental results demonstrate the effectiveness of our method in reconstructing dynamic 3D scenes with 5 input images in one state. Knowledge NeRF is a new pipeline and promising solution for novel view synthesis in dynamic articulated objects. The data and implementation are publicly available at https://github.com/RussRobin/Knowledge_NeRF.

false

true

false

443,057

2108.06084

The Stability-Efficiency Dilemma: Investigating Sequence Length Warmup for Training GPT Models

Recent works have demonstrated great success in pre-training large-scale autoregressive language models on massive GPUs. To reduce the wall-clock training time, a common practice is to increase the batch size and learning rate. However, such practice is often brittle and leads to a so-called stability-efficiency dilemma: increasing the batch sizes and learning rates leads to better training efficiency but can also result in training instability, leading to poor generalization accuracy or failed runs. To better understand this phenomenon, we conduct an in-depth analysis on large-scale pre-training experiments replicating the GPT-2 model. We find that there is a strong correlation between training instability and extreme values of gradient variance, and that samples with long sequence lengths contribute to these extreme gradient variance values, especially at the beginning of the training, indicating that long sequence length can be a main source of training instability. Based on the analysis, we present a Sequence Length Warmup method that aims to solve the training stability-efficiency dilemma. Experiments replicating GPT-2 models show that our approach enables stable training with 8x larger batch size and 4x larger learning rate, whereas the baseline approach struggles with training instability. To achieve the same or better zero-shot evaluation results, our method reduces the required number of training tokens and wall clock time by up to 2.2x and 3.7x, respectively. Experiments replicating GPT-3 model (125M) show that our approach enables stable training with 8x larger batch size and 40x larger learning rate, and retains 99% of the zero-shot accuracy on 11 tasks using 10x less data and 17x less time compared to the original GPT-3 training recipe, while the baseline diverges under the same settings and only retain 95% of accuracy under lower learning rate.

false

true

false

true

250,497

2305.14263

LIMIT: Language Identification, Misidentification, and Translation using Hierarchical Models in 350+ Languages

Knowing the language of an input text/audio is a necessary first step for using almost every NLP tool such as taggers, parsers, or translation systems. Language identification is a well-studied problem, sometimes even considered solved; in reality, due to lack of data and computational challenges, current systems cannot accurately identify most of the world's 7000 languages. To tackle this bottleneck, we first compile a corpus, MCS-350, of 50K multilingual and parallel children's stories in 350+ languages. MCS-350 can serve as a benchmark for language identification of short texts and for 1400+ new translation directions in low-resource Indian and African languages. Second, we propose a novel misprediction-resolution hierarchical model, LIMIt, for language identification that reduces error by 55% (from 0.71 to 0.32) on our compiled children's stories dataset and by 40% (from 0.23 to 0.14) on the FLORES-200 benchmark. Our method can expand language identification coverage into low-resource languages by relying solely on systemic misprediction patterns, bypassing the need to retrain large models from scratch.

false

true

false

true

false

366,936

2304.09802

Generalization and Estimation Error Bounds for Model-based Neural Networks

Model-based neural networks provide unparalleled performance for various tasks, such as sparse coding and compressed sensing problems. Due to the strong connection with the sensing model, these networks are interpretable and inherit prior structure of the problem. In practice, model-based neural networks exhibit higher generalization capability compared to ReLU neural networks. However, this phenomenon was not addressed theoretically. Here, we leverage complexity measures including the global and local Rademacher complexities, in order to provide upper bounds on the generalization and estimation errors of model-based networks. We show that the generalization abilities of model-based networks for sparse recovery outperform those of regular ReLU networks, and derive practical design rules that allow to construct model-based networks with guaranteed high generalization. We demonstrate through a series of experiments that our theoretical insights shed light on a few behaviours experienced in practice, including the fact that ISTA and ADMM networks exhibit higher generalization abilities (especially for small number of training samples), compared to ReLU networks.

false

true

false

359,177