paper_id
stringlengths
9
24
title
stringlengths
10
192
paper_url
stringlengths
44
59
authors
listlengths
1
23
type
stringclasses
0 values
primary_area
stringclasses
0 values
abstract
large_stringlengths
82
4.49k
keywords
listlengths
0
0
TL;DR
large_stringclasses
0 values
submission_number
int64
1
1.18k
arxiv_id
stringlengths
10
10
arxiv_id_source
stringclasses
2 values
v139/kim21a
Improving Predictors via Combination Across Diverse Task Categories
https://proceedings.mlr.press/v139/kim21a.html
[ "Kwang In Kim" ]
null
null
Predictor combination is the problem of improving a task predictor using predictors of other tasks when the forms of individual predictors are unknown. Previous work approached this problem by nonparametrically assessing predictor relationships based on their joint evaluations on a shared sample. This limits their appl...
[]
null
501
null
null
v139/kim21b
Self-Improved Retrosynthetic Planning
https://proceedings.mlr.press/v139/kim21b.html
[ "Junsu Kim", "Sungsoo Ahn", "Hankook Lee", "Jinwoo Shin" ]
null
null
Retrosynthetic planning is a fundamental problem in chemistry for finding a pathway of reactions to synthesize a target molecule. Recently, search algorithms have shown promising results for solving this problem by using deep neural networks (DNNs) to expand their candidate solutions, i.e., adding new reactions to reac...
[]
null
502
2106.04880
title_snapshot
v139/kim21c
Reward Identification in Inverse Reinforcement Learning
https://proceedings.mlr.press/v139/kim21c.html
[ "Kuno Kim", "Shivam Garg", "Kirankumar Shiragur", "Stefano Ermon" ]
null
null
We study the problem of reward identifiability in the context of Inverse Reinforcement Learning (IRL). The reward identifiability question is critical to answer when reasoning about the effectiveness of using Markov Decision Processes (MDPs) as computational models of real world decision makers in order to understand c...
[]
null
503
null
null
v139/kim21d
I-BERT: Integer-only BERT Quantization
https://proceedings.mlr.press/v139/kim21d.html
[ "Sehoon Kim", "Amir Gholami", "Zhewei Yao", "Michael W. Mahoney", "Kurt Keutzer" ]
null
null
Transformer based models, like BERT and RoBERTa, have achieved state-of-the-art results in many Natural Language Processing tasks. However, their memory footprint, inference latency, and power consumption are prohibitive efficient inference at the edge, and even at the data center. While quantization can be a viable so...
[]
null
504
2101.01321
title_snapshot
v139/kim21e
Message Passing Adaptive Resonance Theory for Online Active Semi-supervised Learning
https://proceedings.mlr.press/v139/kim21e.html
[ "Taehyeong Kim", "Injune Hwang", "Hyundo Lee", "Hyunseo Kim", "Won-Seok Choi", "Joseph J Lim", "Byoung-Tak Zhang" ]
null
null
Active learning is widely used to reduce labeling effort and training time by repeatedly querying only the most beneficial samples from unlabeled data. In real-world problems where data cannot be stored indefinitely due to limited storage or privacy issues, the query selection and the model update should be performed a...
[]
null
505
2012.01227
title_snapshot
v139/kim21f
Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech
https://proceedings.mlr.press/v139/kim21f.html
[ "Jaehyeon Kim", "Jungil Kong", "Juhee Son" ]
null
null
Several recent end-to-end text-to-speech (TTS) models enabling single-stage training and parallel sampling have been proposed, but their sample quality does not match that of two-stage TTS systems. In this work, we present a parallel end-to-end TTS method that generates more natural sounding audio than current two-stag...
[]
null
506
2106.06103
title_snapshot
v139/kim21g
A Policy Gradient Algorithm for Learning to Learn in Multiagent Reinforcement Learning
https://proceedings.mlr.press/v139/kim21g.html
[ "Dong Ki Kim", "Miao Liu", "Matthew D Riemer", "Chuangchuang Sun", "Marwa Abdulhai", "Golnaz Habibi", "Sebastian Lopez-Cot", "Gerald Tesauro", "Jonathan How" ]
null
null
A fundamental challenge in multiagent reinforcement learning is to learn beneficial behaviors in a shared environment with other simultaneously learning agents. In particular, each agent perceives the environment as effectively non-stationary due to the changing policies of other agents. Moreover, each agent is itself ...
[]
null
507
2011.00382
title_snapshot
v139/kim21h
Inferring Latent Dynamics Underlying Neural Population Activity via Neural Differential Equations
https://proceedings.mlr.press/v139/kim21h.html
[ "Timothy D. Kim", "Thomas Z. Luo", "Jonathan W. Pillow", "Carlos D. Brody" ]
null
null
An important problem in systems neuroscience is to identify the latent dynamics underlying neural population activity. Here we address this problem by introducing a low-dimensional nonlinear model for latent neural population dynamics using neural ordinary differential equations (neural ODEs), with noisy sensory inputs...
[]
null
508
null
null
v139/kim21i
The Lipschitz Constant of Self-Attention
https://proceedings.mlr.press/v139/kim21i.html
[ "Hyunjik Kim", "George Papamakarios", "Andriy Mnih" ]
null
null
Lipschitz constants of neural networks have been explored in various contexts in deep learning, such as provable adversarial robustness, estimating Wasserstein distance, stabilising training of GANs, and formulating invertible neural networks. Such works have focused on bounding the Lipschitz constant of fully connecte...
[]
null
509
2006.04710
title_snapshot
v139/kim21j
Unsupervised Skill Discovery with Bottleneck Option Learning
https://proceedings.mlr.press/v139/kim21j.html
[ "Jaekyeom Kim", "Seohong Park", "Gunhee Kim" ]
null
null
Having the ability to acquire inherent skills from environments without any external rewards or supervision like humans is an important problem. We propose a novel unsupervised skill discovery method named Information Bottleneck Option Learning (IBOL). On top of the linearization of environments that promotes more vari...
[]
null
510
2106.14305
title_snapshot
v139/kim21k
ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision
https://proceedings.mlr.press/v139/kim21k.html
[ "Wonjae Kim", "Bokyung Son", "Ildoo Kim" ]
null
null
Vision-and-Language Pre-training (VLP) has improved performance on various joint vision-and-language downstream tasks. Current approaches to VLP heavily rely on image feature extraction processes, most of which involve region supervision (e.g., object detection) and the convolutional architecture (e.g., ResNet). Althou...
[]
null
511
2102.03334
title_snapshot
v139/kirschner21a
Bias-Robust Bayesian Optimization via Dueling Bandits
https://proceedings.mlr.press/v139/kirschner21a.html
[ "Johannes Kirschner", "Andreas Krause" ]
null
null
We consider Bayesian optimization in settings where observations can be adversarially biased, for example by an uncontrolled hidden confounder. Our first contribution is a reduction of the confounded setting to the dueling bandit model. Then we propose a novel approach for dueling bandits based on information-directed ...
[]
null
512
2105.11802
title_snapshot
v139/kiyasseh21a
CLOCS: Contrastive Learning of Cardiac Signals Across Space, Time, and Patients
https://proceedings.mlr.press/v139/kiyasseh21a.html
[ "Dani Kiyasseh", "Tingting Zhu", "David A Clifton" ]
null
null
The healthcare industry generates troves of unlabelled physiological data. This data can be exploited via contrastive learning, a self-supervised pre-training method that encourages representations of instances to be similar to one another. We propose a family of contrastive learning methods, CLOCS, that encourages rep...
[]
null
513
2005.13249
title_snapshot
v139/gasteiger21a
Scalable Optimal Transport in High Dimensions for Graph Distances, Embedding Alignment, and More
https://proceedings.mlr.press/v139/gasteiger21a.html
[ "Johannes Gasteiger", "Marten Lienen", "Stephan Günnemann" ]
null
null
The current best practice for computing optimal transport (OT) is via entropy regularization and Sinkhorn iterations. This algorithm runs in quadratic time as it requires the full pairwise cost matrix, which is prohibitively expensive for large sets of objects. In this work we propose two effective log-linear time appr...
[]
null
514
2107.06876
title_snapshot
v139/koehler21a
Representational aspects of depth and conditioning in normalizing flows
https://proceedings.mlr.press/v139/koehler21a.html
[ "Frederic Koehler", "Viraj Mehta", "Andrej Risteski" ]
null
null
Normalizing flows are among the most popular paradigms in generative modeling, especially for images, primarily because we can efficiently evaluate the likelihood of a data point. This is desirable both for evaluating the fit of a model, and for ease of training, as maximizing the likelihood can be done by gradient des...
[]
null
515
2010.01155
title_snapshot
v139/koh21a
WILDS: A Benchmark of in-the-Wild Distribution Shifts
https://proceedings.mlr.press/v139/koh21a.html
[ "Pang Wei Koh", "Shiori Sagawa", "Henrik Marklund", "Sang Michael Xie", "Marvin Zhang", "Akshay Balsubramani", "Weihua Hu", "Michihiro Yasunaga", "Richard Lanas Phillips", "Irena Gao", "Tony Lee", "Etienne David", "Ian Stavness", "Wei Guo", "Berton Earnshaw", "Imran Haque", "Sara M B...
null
null
Distribution shifts—where the training distribution differs from the test distribution—can substantially degrade the accuracy of machine learning (ML) systems deployed in the wild. Despite their ubiquity in the real-world deployments, these distribution shifts are under-represented in the datasets widely used in the ML...
[]
null
516
2012.07421
title_snapshot
v139/kolmogorov21a
One-sided Frank-Wolfe algorithms for saddle problems
https://proceedings.mlr.press/v139/kolmogorov21a.html
[ "Vladimir Kolmogorov", "Thomas Pock" ]
null
null
We study a class of convex-concave saddle-point problems of the form $\min_x\max_y ⟨Kx,y⟩+f_{\cal P}(x)-h^*(y)$ where $K$ is a linear operator, $f_{\cal P}$ is the sum of a convex function $f$ with a Lipschitz-continuous gradient and the indicator function of a bounded convex polytope ${\cal P}$, and $h^\ast$ is a conv...
[]
null
517
2101.12617
title_snapshot
v139/komanduru21a
A Lower Bound for the Sample Complexity of Inverse Reinforcement Learning
https://proceedings.mlr.press/v139/komanduru21a.html
[ "Abi Komanduru", "Jean Honorio" ]
null
null
Inverse reinforcement learning (IRL) is the task of finding a reward function that generates a desired optimal policy for a given Markov Decision Process (MDP). This paper develops an information-theoretic lower bound for the sample complexity of the finite state, finite action IRL problem. A geometric construction of ...
[]
null
518
2103.04446
title_snapshot
v139/kong21a
Consensus Control for Decentralized Deep Learning
https://proceedings.mlr.press/v139/kong21a.html
[ "Lingjing Kong", "Tao Lin", "Anastasia Koloskova", "Martin Jaggi", "Sebastian Stich" ]
null
null
Decentralized training of deep learning models enables on-device learning over networks, as well as efficient scaling to large compute clusters. Experiments in earlier works reveal that, even in a data-center setup, decentralized training often suffers from the degradation in the quality of the model: the training and ...
[]
null
519
2102.04828
title_snapshot
v139/konobeev21a
A Distribution-dependent Analysis of Meta Learning
https://proceedings.mlr.press/v139/konobeev21a.html
[ "Mikhail Konobeev", "Ilja Kuzborskij", "Csaba Szepesvari" ]
null
null
A key problem in the theory of meta-learning is to understand how the task distributions influence transfer risk, the expected error of a meta-learner on a new task drawn from the unknown task distribution. In this paper, focusing on fixed design linear regression with Gaussian noise and a Gaussian task (or parameter) ...
[]
null
520
2011.00344
title_snapshot
v139/kopetzki21a
Evaluating Robustness of Predictive Uncertainty Estimation: Are Dirichlet-based Models Reliable?
https://proceedings.mlr.press/v139/kopetzki21a.html
[ "Anna-Kathrin Kopetzki", "Bertrand Charpentier", "Daniel Zügner", "Sandhya Giri", "Stephan Günnemann" ]
null
null
Dirichlet-based uncertainty (DBU) models are a recent and promising class of uncertainty-aware models. DBU models predict the parameters of a Dirichlet distribution to provide fast, high-quality uncertainty estimates alongside with class predictions. In this work, we present the first large-scale, in-depth study of the...
[]
null
521
2010.14986
title_snapshot
v139/korba21a
Kernel Stein Discrepancy Descent
https://proceedings.mlr.press/v139/korba21a.html
[ "Anna Korba", "Pierre-Cyril Aubin-Frankowski", "Szymon Majewski", "Pierre Ablin" ]
null
null
Among dissimilarities between probability distributions, the Kernel Stein Discrepancy (KSD) has received much interest recently. We investigate the properties of its Wasserstein gradient flow to approximate a target probability distribution $\pi$ on $\mathbb{R}^d$, known up to a normalization constant. This leads to a ...
[]
null
522
2105.09994
title_snapshot
v139/kosaian21a
Boosting the Throughput and Accelerator Utilization of Specialized CNN Inference Beyond Increasing Batch Size
https://proceedings.mlr.press/v139/kosaian21a.html
[ "Jack Kosaian", "Amar Phanishayee", "Matthai Philipose", "Debadeepta Dey", "Rashmi Vinayak" ]
null
null
Datacenter vision systems widely use small, specialized convolutional neural networks (CNNs) trained on specific tasks for high-throughput inference. These settings employ accelerators with massive computational capacity, but which specialized CNNs underutilize due to having low arithmetic intensity. This results in su...
[]
null
523
null
null
v139/kosiorek21a
NeRF-VAE: A Geometry Aware 3D Scene Generative Model
https://proceedings.mlr.press/v139/kosiorek21a.html
[ "Adam R Kosiorek", "Heiko Strathmann", "Daniel Zoran", "Pol Moreno", "Rosalia Schneider", "Sona Mokra", "Danilo Jimenez Rezende" ]
null
null
We propose NeRF-VAE, a 3D scene generative model that incorporates geometric structure via Neural Radiance Fields (NeRF) and differentiable volume rendering. In contrast to NeRF, our model takes into account shared structure across scenes, and is able to infer the structure of a novel scene—without the need to re-train...
[]
null
524
2104.00587
title_snapshot
v139/kossen21a
Active Testing: Sample-Efficient Model Evaluation
https://proceedings.mlr.press/v139/kossen21a.html
[ "Jannik Kossen", "Sebastian Farquhar", "Yarin Gal", "Tom Rainforth" ]
null
null
We introduce a new framework for sample-efficient model evaluation that we call active testing. While approaches like active learning reduce the number of labels needed for model training, existing literature largely ignores the cost of labeling test data, typically unrealistically assuming large test sets for model ev...
[]
null
525
2103.05331
title_snapshot
v139/kostas21a
High Confidence Generalization for Reinforcement Learning
https://proceedings.mlr.press/v139/kostas21a.html
[ "James Kostas", "Yash Chandak", "Scott M Jordan", "Georgios Theocharous", "Philip Thomas" ]
null
null
We present several classes of reinforcement learning algorithms that safely generalize to Markov decision processes (MDPs) not seen during training. Specifically, we study the setting in which some set of MDPs is accessible for training. The goal is to generalize safely to MDPs that are sampled from the same distributi...
[]
null
526
null
null
v139/kostrikov21a
Offline Reinforcement Learning with Fisher Divergence Critic Regularization
https://proceedings.mlr.press/v139/kostrikov21a.html
[ "Ilya Kostrikov", "Rob Fergus", "Jonathan Tompson", "Ofir Nachum" ]
null
null
Many modern approaches to offline Reinforcement Learning (RL) utilize behavior regularization, typically augmenting a model-free actor critic algorithm with a penalty measuring divergence of the policy from the offline data. In this work, we propose an alternative approach to encouraging the learned policy to stay clos...
[]
null
527
2103.08050
title_snapshot
v139/kovalev21a
ADOM: Accelerated Decentralized Optimization Method for Time-Varying Networks
https://proceedings.mlr.press/v139/kovalev21a.html
[ "Dmitry Kovalev", "Egor Shulgin", "Peter Richtarik", "Alexander V Rogozin", "Alexander Gasnikov" ]
null
null
We propose ADOM – an accelerated method for smooth and strongly convex decentralized optimization over time-varying networks. ADOM uses a dual oracle, i.e., we assume access to the gradient of the Fenchel conjugate of the individual loss functions. Up to a constant factor, which depends on the network structure only, i...
[]
null
528
2102.09234
title_snapshot
v139/kozuno21a
Revisiting Peng’s Q($λ$) for Modern Reinforcement Learning
https://proceedings.mlr.press/v139/kozuno21a.html
[ "Tadashi Kozuno", "Yunhao Tang", "Mark Rowland", "Remi Munos", "Steven Kapturowski", "Will Dabney", "Michal Valko", "David Abel" ]
null
null
Off-policy multi-step reinforcement learning algorithms consist of conservative and non-conservative algorithms: the former actively cut traces, whereas the latter do not. Recently, Munos et al. (2016) proved the convergence of conservative algorithms to an optimal Q-function. In contrast, non-conservative algorithms a...
[]
null
529
2103.00107
title_snapshot
v139/krishnamurthy21a
Adapting to misspecification in contextual bandits with offline regression oracles
https://proceedings.mlr.press/v139/krishnamurthy21a.html
[ "Sanath Kumar Krishnamurthy", "Vitor Hadad", "Susan Athey" ]
null
null
Computationally efficient contextual bandits are often based on estimating a predictive model of rewards given contexts and arms using past data. However, when the reward model is not well-specified, the bandit algorithm may incur unexpected regret, so recent work has focused on algorithms that are robust to misspecifi...
[]
null
530
2102.13240
title_snapshot
v139/krueger21a
Out-of-Distribution Generalization via Risk Extrapolation (REx)
https://proceedings.mlr.press/v139/krueger21a.html
[ "David Krueger", "Ethan Caballero", "Joern-Henrik Jacobsen", "Amy Zhang", "Jonathan Binas", "Dinghuai Zhang", "Remi Le Priol", "Aaron Courville" ]
null
null
Distributional shift is one of the major obstacles when transferring machine learning prediction systems from the lab to the real world. To tackle this problem, we assume that variation across training domains is representative of the variation we might encounter at test time, but also that shifts at test time may be m...
[]
null
531
2003.00688
title_snapshot
v139/kuchibhotla21a
Near-Optimal Confidence Sequences for Bounded Random Variables
https://proceedings.mlr.press/v139/kuchibhotla21a.html
[ "Arun K Kuchibhotla", "Qinqing Zheng" ]
null
null
Many inference problems, such as sequential decision problems like A/B testing, adaptive sampling schemes like bandit selection, are often online in nature. The fundamental problem for online inference is to provide a sequence of confidence intervals that are valid uniformly over the growing-into-infinity sample sizes....
[]
null
532
2006.05022
title_snapshot
v139/kulkarni21a
Differentially Private Bayesian Inference for Generalized Linear Models
https://proceedings.mlr.press/v139/kulkarni21a.html
[ "Tejas Kulkarni", "Joonas Jälkö", "Antti Koskela", "Samuel Kaski", "Antti Honkela" ]
null
null
Generalized linear models (GLMs) such as logistic regression are among the most widely used arms in data analyst’s repertoire and often used on sensitive datasets. A large body of prior works that investigate GLMs under differential privacy (DP) constraints provide only private point estimates of the regression coeffic...
[]
null
533
2011.00467
title_snapshot
v139/kumar21a
Bayesian Structural Adaptation for Continual Learning
https://proceedings.mlr.press/v139/kumar21a.html
[ "Abhishek Kumar", "Sunabha Chatterjee", "Piyush Rai" ]
null
null
Continual Learning is a learning paradigm where learning systems are trained on a sequence of tasks. The goal here is to perform well on the current task without suffering from a performance drop on the previous tasks. Two notable directions among the recent advances in continual learning with neural networks are (1) v...
[]
null
534
null
null
v139/kumar21b
Implicit rate-constrained optimization of non-decomposable objectives
https://proceedings.mlr.press/v139/kumar21b.html
[ "Abhishek Kumar", "Harikrishna Narasimhan", "Andrew Cotter" ]
null
null
We consider a popular family of constrained optimization problems arising in machine learning that involve optimizing a non-decomposable evaluation metric with a certain thresholded form, while constraining another metric of interest. Examples of such problems include optimizing false negative rate at a fixed false pos...
[]
null
535
2107.10960
title_snapshot
v139/kummerle21a
A Scalable Second Order Method for Ill-Conditioned Matrix Completion from Few Samples
https://proceedings.mlr.press/v139/kummerle21a.html
[ "Christian Kümmerle", "Claudio M. Verdun" ]
null
null
We propose an iterative algorithm for low-rank matrix completion with that can be interpreted as an iteratively reweighted least squares (IRLS) algorithm, a saddle-escaping smoothing Newton method or a variable metric proximal gradient method applied to a non-convex rank surrogate. It combines the favorable data-effici...
[]
null
536
2106.02119
title_snapshot
v139/kveton21a
Meta-Thompson Sampling
https://proceedings.mlr.press/v139/kveton21a.html
[ "Branislav Kveton", "Mikhail Konobeev", "Manzil Zaheer", "Chih-Wei Hsu", "Martin Mladenov", "Craig Boutilier", "Csaba Szepesvari" ]
null
null
Efficient exploration in bandits is a fundamental online learning problem. We propose a variant of Thompson sampling that learns to explore better as it interacts with bandit instances drawn from an unknown prior. The algorithm meta-learns the prior and thus we call it MetaTS. We propose several efficient implementatio...
[]
null
537
2102.06129
title_snapshot
v139/kwon21a
Targeted Data Acquisition for Evolving Negotiation Agents
https://proceedings.mlr.press/v139/kwon21a.html
[ "Minae Kwon", "Siddharth Karamcheti", "Mariano-Florentino Cuellar", "Dorsa Sadigh" ]
null
null
Successful negotiators must learn how to balance optimizing for self-interest and cooperation. Yet current artificial negotiation agents often heavily depend on the quality of the static datasets they were trained on, limiting their capacity to fashion an adaptive response balancing self-interest and cooperation. For t...
[]
null
538
2106.07728
title_snapshot
v139/kwon21b
ASAM: Adaptive Sharpness-Aware Minimization for Scale-Invariant Learning of Deep Neural Networks
https://proceedings.mlr.press/v139/kwon21b.html
[ "Jungmin Kwon", "Jeongseop Kim", "Hyunseo Park", "In Kwon Choi" ]
null
null
Recently, learning algorithms motivated from sharpness of loss surface as an effective measure of generalization gap have shown state-of-the-art performances. Nevertheless, sharpness defined in a rigid region with a fixed radius, has a drawback in sensitivity to parameter re-scaling which leaves the loss unaffected, le...
[]
null
539
2102.11600
title_snapshot
v139/laber21a
On the price of explainability for some clustering problems
https://proceedings.mlr.press/v139/laber21a.html
[ "Eduardo S Laber", "Lucas Murtinho" ]
null
null
The price of explainability for a clustering task can be defined as the unavoidable loss, in terms of the objective function, if we force the final partition to be explainable. Here, we study this price for the following clustering problems: $k$-means, $k$-medians, $k$-centers and maximum-spacing. We provide upper and ...
[]
null
540
2101.01576
title_snapshot
v139/lacotte21a
Adaptive Newton Sketch: Linear-time Optimization with Quadratic Convergence and Effective Hessian Dimensionality
https://proceedings.mlr.press/v139/lacotte21a.html
[ "Jonathan Lacotte", "Yifei Wang", "Mert Pilanci" ]
null
null
We propose a randomized algorithm with quadratic convergence rate for convex optimization problems with a self-concordant, composite, strongly convex objective function. Our method is based on performing an approximate Newton step using a random projection of the Hessian. Our first contribution is to show that, at each...
[]
null
541
2105.07291
title_snapshot
v139/laforgue21a
Generalization Bounds in the Presence of Outliers: a Median-of-Means Study
https://proceedings.mlr.press/v139/laforgue21a.html
[ "Pierre Laforgue", "Guillaume Staerman", "Stephan Clémençon" ]
null
null
In contrast to the empirical mean, the Median-of-Means (MoM) is an estimator of the mean $\theta$ of a square integrable r.v. Z, around which accurate nonasymptotic confidence bounds can be built, even when Z does not exhibit a sub-Gaussian tail behavior. Thanks to the high confidence it achieves on heavy-tailed data, ...
[]
null
542
2006.05240
title_snapshot
v139/lam21a
Model Fusion for Personalized Learning
https://proceedings.mlr.press/v139/lam21a.html
[ "Thanh Chi Lam", "Nghia Hoang", "Bryan Kian Hsiang Low", "Patrick Jaillet" ]
null
null
Production systems operating on a growing domain of analytic services often require generating warm-start solution models for emerging tasks with limited data. One potential approach to address this warm-start challenge is to adopt meta learning to generate a base model that can be adapted to solve unseen tasks with mi...
[]
null
543
null
null
v139/lam21b
Gradient Disaggregation: Breaking Privacy in Federated Learning by Reconstructing the User Participant Matrix
https://proceedings.mlr.press/v139/lam21b.html
[ "Maximilian Lam", "Gu-Yeon Wei", "David Brooks", "Vijay Janapa Reddi", "Michael Mitzenmacher" ]
null
null
We show that aggregated model updates in federated learning may be insecure. An untrusted central server may disaggregate user updates from sums of updates across participants given repeated observations, enabling the server to recover privileged information about individual users’ private training data via traditional...
[]
null
544
2106.06089
title_snapshot
v139/lancewicki21a
Stochastic Multi-Armed Bandits with Unrestricted Delay Distributions
https://proceedings.mlr.press/v139/lancewicki21a.html
[ "Tal Lancewicki", "Shahar Segal", "Tomer Koren", "Yishay Mansour" ]
null
null
We study the stochastic Multi-Armed Bandit (MAB) problem with random delays in the feedback received by the algorithm. We consider two settings: the {\it reward dependent} delay setting, where realized delays may depend on the stochastic rewards, and the {\it reward-independent} delay setting. Our main contribution is ...
[]
null
545
2106.02436
title_snapshot
v139/landajuela21a
Discovering symbolic policies with deep reinforcement learning
https://proceedings.mlr.press/v139/landajuela21a.html
[ "Mikel Landajuela", "Brenden K Petersen", "Sookyung Kim", "Claudio P Santiago", "Ruben Glatt", "Nathan Mundhenk", "Jacob F Pettit", "Daniel Faissol" ]
null
null
Deep reinforcement learning (DRL) has proven successful for many difficult control problems by learning policies represented by neural networks. However, the complexity of neural network-based policies{—}involving thousands of composed non-linear operators{—}can render them problematic to understand, trust, and deploy....
[]
null
546
null
null
v139/lang21a
Graph Cuts Always Find a Global Optimum for Potts Models (With a Catch)
https://proceedings.mlr.press/v139/lang21a.html
[ "Hunter Lang", "David Sontag", "Aravindan Vijayaraghavan" ]
null
null
We prove that the alpha-expansion algorithm for MAP inference always returns a globally optimal assignment for Markov Random Fields with Potts pairwise potentials, with a catch: the returned assignment is only guaranteed to be optimal for an instance within a small perturbation of the original problem instance. In othe...
[]
null
547
2011.03639
title_snapshot
v139/lange21a
Efficient Message Passing for 0–1 ILPs with Binary Decision Diagrams
https://proceedings.mlr.press/v139/lange21a.html
[ "Jan-Hendrik Lange", "Paul Swoboda" ]
null
null
We present a message passing method for 0{–}1 integer linear programs. Our algorithm is based on a decomposition of the original problem into subproblems that are represented as binary deci- sion diagrams. The resulting Lagrangean dual is solved iteratively by a series of efficient block coordinate ascent steps. Our me...
[]
null
548
2009.00481
title_snapshot
v139/larsen21a
CountSketches, Feature Hashing and the Median of Three
https://proceedings.mlr.press/v139/larsen21a.html
[ "Kasper Green Larsen", "Rasmus Pagh", "Jakub Tětek" ]
null
null
In this paper, we revisit the classic CountSketch method, which is a sparse, random projection that transforms a (high-dimensional) Euclidean vector $v$ to a vector of dimension $(2t-1) s$, where $t, s > 0$ are integer parameters. It is known that a CountSketch allows estimating coordinates of $v$ with variance bounded...
[]
null
549
2102.02193
title_snapshot
v139/laturnus21a
MorphVAE: Generating Neural Morphologies from 3D-Walks using a Variational Autoencoder with Spherical Latent Space
https://proceedings.mlr.press/v139/laturnus21a.html
[ "Sophie C. Laturnus", "Philipp Berens" ]
null
null
For the past century, the anatomy of a neuron has been considered one of its defining features: The shape of a neuron’s dendrites and axon fundamentally determines what other neurons it can connect to. These neurites have been described using mathematical tools e.g. in the context of cell type classification, but gener...
[]
null
550
null
null
v139/lazic21a
Improved Regret Bound and Experience Replay in Regularized Policy Iteration
https://proceedings.mlr.press/v139/lazic21a.html
[ "Nevena Lazic", "Dong Yin", "Yasin Abbasi-Yadkori", "Csaba Szepesvari" ]
null
null
In this work, we study algorithms for learning in infinite-horizon undiscounted Markov decision processes (MDPs) with function approximation. We first show that the regret analysis of the Politex algorithm (a version of regularized policy iteration) can be sharpened from $O(T^{3/4})$ to $O(\sqrt{T})$ under nearly ident...
[]
null
551
2102.12611
title_snapshot
v139/le21a
LAMDA: Label Matching Deep Domain Adaptation
https://proceedings.mlr.press/v139/le21a.html
[ "Trung Le", "Tuan Nguyen", "Nhat Ho", "Hung Bui", "Dinh Phung" ]
null
null
Deep domain adaptation (DDA) approaches have recently been shown to perform better than their shallow rivals with better modeling capacity on complex domains (e.g., image, structural data, and sequential data). The underlying idea is to learn domain invariant representations on a latent space that can bridge the gap be...
[]
null
552
null
null
v139/lederer21a
Gaussian Process-Based Real-Time Learning for Safety Critical Applications
https://proceedings.mlr.press/v139/lederer21a.html
[ "Armin Lederer", "Alejandro J Ordóñez Conejo", "Korbinian A Maier", "Wenxin Xiao", "Jonas Umlauft", "Sandra Hirche" ]
null
null
The safe operation of physical systems typically relies on high-quality models. Since a continuous stream of data is generated during run-time, such models are often obtained through the application of Gaussian process regression because it provides guarantees on the prediction error. Due to its high computational comp...
[]
null
553
null
null
v139/lee21a
Sharing Less is More: Lifelong Learning in Deep Networks with Selective Layer Transfer
https://proceedings.mlr.press/v139/lee21a.html
[ "Seungwon Lee", "Sima Behpour", "Eric Eaton" ]
null
null
Effective lifelong learning across diverse tasks requires the transfer of diverse knowledge, yet transferring irrelevant knowledge may lead to interference and catastrophic forgetting. In deep networks, transferring the appropriate granularity of knowledge is as important as the transfer mechanism, and must be driven b...
[]
null
554
null
null
v139/lee21b
Fair Selective Classification Via Sufficiency
https://proceedings.mlr.press/v139/lee21b.html
[ "Joshua K Lee", "Yuheng Bu", "Deepta Rajan", "Prasanna Sattigeri", "Rameswar Panda", "Subhro Das", "Gregory W Wornell" ]
null
null
Selective classification is a powerful tool for decision-making in scenarios where mistakes are costly but abstentions are allowed. In general, by allowing a classifier to abstain, one can improve the performance of a model at the cost of reducing coverage and classifying fewer samples. However, recent work has shown, ...
[]
null
555
null
null
v139/lee21c
On-the-fly Rectification for Robust Large-Vocabulary Topic Inference
https://proceedings.mlr.press/v139/lee21c.html
[ "Moontae Lee", "Sungjun Cho", "Kun Dong", "David Mimno", "David Bindel" ]
null
null
Across many data domains, co-occurrence statistics about the joint appearance of objects are powerfully informative. By transforming unsupervised learning problems into decompositions of co-occurrence statistics, spectral algorithms provide transparent and efficient algorithms for posterior inference such as latent top...
[]
null
556
2111.06580
title_snapshot
v139/lee21d
Unsupervised Embedding Adaptation via Early-Stage Feature Reconstruction for Few-Shot Classification
https://proceedings.mlr.press/v139/lee21d.html
[ "Dong Hoon Lee", "Sae-Young Chung" ]
null
null
We propose unsupervised embedding adaptation for the downstream few-shot classification task. Based on findings that deep neural networks learn to generalize before memorizing, we develop Early-Stage Feature Reconstruction (ESFR) — a novel adaptation scheme with feature reconstruction and dimensionality-driven early st...
[]
null
557
2106.11486
title_snapshot
v139/lee21e
Continual Learning in the Teacher-Student Setup: Impact of Task Similarity
https://proceedings.mlr.press/v139/lee21e.html
[ "Sebastian Lee", "Sebastian Goldt", "Andrew Saxe" ]
null
null
Continual learning{—}the ability to learn many tasks in sequence{—}is critical for artificial learning systems. Yet standard training methods for deep networks often suffer from catastrophic forgetting, where learning new tasks erases knowledge of the earlier tasks. While catastrophic forgetting labels the problem, the...
[]
null
558
2107.04384
title_snapshot
v139/lee21f
OptiDICE: Offline Policy Optimization via Stationary Distribution Correction Estimation
https://proceedings.mlr.press/v139/lee21f.html
[ "Jongmin Lee", "Wonseok Jeon", "Byungjun Lee", "Joelle Pineau", "Kee-Eung Kim" ]
null
null
We consider the offline reinforcement learning (RL) setting where the agent aims to optimize the policy solely from the data without further environment interactions. In offline RL, the distributional shift becomes the primary source of difficulty, which arises from the deviation of the target policy being optimized fr...
[]
null
559
2106.10783
title_snapshot
v139/lee21g
SUNRISE: A Simple Unified Framework for Ensemble Learning in Deep Reinforcement Learning
https://proceedings.mlr.press/v139/lee21g.html
[ "Kimin Lee", "Michael Laskin", "Aravind Srinivas", "Pieter Abbeel" ]
null
null
Off-policy deep reinforcement learning (RL) has been successful in a range of challenging domains. However, standard off-policy RL algorithms can suffer from several issues, such as instability in Q-learning and balancing exploration and exploitation. To mitigate these issues, we present SUNRISE, a simple unified ensem...
[]
null
560
2007.04938
title_snapshot
v139/lee21h
Achieving Near Instance-Optimality and Minimax-Optimality in Stochastic and Adversarial Linear Bandits Simultaneously
https://proceedings.mlr.press/v139/lee21h.html
[ "Chung-Wei Lee", "Haipeng Luo", "Chen-Yu Wei", "Mengxiao Zhang", "Xiaojin Zhang" ]
null
null
In this work, we develop linear bandit algorithms that automatically adapt to different environments. By plugging a novel loss estimator into the optimization problem that characterizes the instance-optimal strategy, our first algorithm not only achieves nearly instance-optimal regret in stochastic environments, but al...
[]
null
561
2102.05858
title_snapshot
v139/lee21i
PEBBLE: Feedback-Efficient Interactive Reinforcement Learning via Relabeling Experience and Unsupervised Pre-training
https://proceedings.mlr.press/v139/lee21i.html
[ "Kimin Lee", "Laura M Smith", "Pieter Abbeel" ]
null
null
Conveying complex objectives to reinforcement learning (RL) agents can often be difficult, involving meticulous design of reward functions that are sufficiently informative yet easy enough to provide. Human-in-the-loop RL methods allow practitioners to instead interactively teach agents through tailored feedback; howev...
[]
null
562
2106.05091
title_snapshot
v139/lei21a
Near-Optimal Linear Regression under Distribution Shift
https://proceedings.mlr.press/v139/lei21a.html
[ "Qi Lei", "Wei Hu", "Jason Lee" ]
null
null
Transfer learning is essential when sufficient data comes from the source domain, with scarce labeled data from the target domain. We develop estimators that achieve minimax linear risk for linear regression problems under distribution shift. Our algorithms cover different transfer learning settings including covariate...
[]
null
563
2106.12108
title_snapshot
v139/lei21b
Stability and Generalization of Stochastic Gradient Methods for Minimax Problems
https://proceedings.mlr.press/v139/lei21b.html
[ "Yunwen Lei", "Zhenhuan Yang", "Tianbao Yang", "Yiming Ying" ]
null
null
Many machine learning problems can be formulated as minimax problems such as Generative Adversarial Networks (GANs), AUC maximization and robust estimation, to mention but a few. A substantial amount of studies are devoted to studying the convergence behavior of their stochastic gradient-type algorithms. In contrast, t...
[]
null
564
2105.03793
title_snapshot
v139/leibo21a
Scalable Evaluation of Multi-Agent Reinforcement Learning with Melting Pot
https://proceedings.mlr.press/v139/leibo21a.html
[ "Joel Z Leibo", "Edgar A Dueñez-Guzman", "Alexander Vezhnevets", "John P Agapiou", "Peter Sunehag", "Raphael Koster", "Jayd Matyas", "Charlie Beattie", "Igor Mordatch", "Thore Graepel" ]
null
null
Existing evaluation suites for multi-agent reinforcement learning (MARL) do not assess generalization to novel situations as their primary objective (unlike supervised learning benchmarks). Our contribution, Melting Pot, is a MARL evaluation suite that fills this gap and uses reinforcement learning to reduce the human ...
[]
null
565
2107.06857
title_snapshot
v139/leimkuhler21a
Better Training using Weight-Constrained Stochastic Dynamics
https://proceedings.mlr.press/v139/leimkuhler21a.html
[ "Benedict Leimkuhler", "Tiffany J Vlaar", "Timothée Pouchon", "Amos Storkey" ]
null
null
We employ constraints to control the parameter space of deep neural networks throughout training. The use of customised, appropriately designed constraints can reduce the vanishing/exploding gradients problem, improve smoothness of classification boundaries, control weight magnitudes and stabilize deep neural networks,...
[]
null
566
2106.10704
title_snapshot
v139/leino21a
Globally-Robust Neural Networks
https://proceedings.mlr.press/v139/leino21a.html
[ "Klas Leino", "Zifan Wang", "Matt Fredrikson" ]
null
null
The threat of adversarial examples has motivated work on training certifiably robust neural networks to facilitate efficient verification of local robustness at inference time. We formalize a notion of global robustness, which captures the operational properties of on-line local robustness certification while yielding ...
[]
null
567
2102.08452
title_snapshot
v139/leme21a
Learning to Price Against a Moving Target
https://proceedings.mlr.press/v139/leme21a.html
[ "Renato Paes Leme", "Balasubramanian Sivan", "Yifeng Teng", "Pratik Worah" ]
null
null
In the Learning to Price setting, a seller posts prices over time with the goal of maximizing revenue while learning the buyer’s valuation. This problem is very well understood when values are stationary (fixed or iid). Here we study the problem where the buyer’s value is a moving target, i.e., they change over time ei...
[]
null
568
2106.04689
title_snapshot
v139/lemercier21a
SigGPDE: Scaling Sparse Gaussian Processes on Sequential Data
https://proceedings.mlr.press/v139/lemercier21a.html
[ "Maud Lemercier", "Cristopher Salvi", "Thomas Cass", "Edwin V. Bonilla", "Theodoros Damoulas", "Terry J Lyons" ]
null
null
Making predictions and quantifying their uncertainty when the input data is sequential is a fundamental learning challenge, recently attracting increasing attention. We develop SigGPDE, a new scalable sparse variational inference framework for Gaussian Processes (GPs) on sequential data. Our contribution is twofold. Fi...
[]
null
569
2105.04211
title_snapshot
v139/levanon21a
Strategic Classification Made Practical
https://proceedings.mlr.press/v139/levanon21a.html
[ "Sagi Levanon", "Nir Rosenfeld" ]
null
null
Strategic classification regards the problem of learning in settings where users can strategically modify their features to improve outcomes. This setting applies broadly, and has received much recent attention. But despite its practical significance, work in this space has so far been predominantly theoretical. In thi...
[]
null
570
2103.01826
title_snapshot
v139/levine21a
Improved, Deterministic Smoothing for L_1 Certified Robustness
https://proceedings.mlr.press/v139/levine21a.html
[ "Alexander J Levine", "Soheil Feizi" ]
null
null
Randomized smoothing is a general technique for computing sample-dependent robustness guarantees against adversarial attacks for deep classifiers. Prior works on randomized smoothing against L_1 adversarial attacks use additive smoothing noise and provide probabilistic robustness guarantees. In this work, we propose a ...
[]
null
571
2103.10834
title_snapshot
v139/lewis21a
BASE Layers: Simplifying Training of Large, Sparse Models
https://proceedings.mlr.press/v139/lewis21a.html
[ "Mike Lewis", "Shruti Bhosale", "Tim Dettmers", "Naman Goyal", "Luke Zettlemoyer" ]
null
null
We introduce a new balanced assignment of experts (BASE) layer for large language models that greatly simplifies existing high capacity sparse layers. Sparse layers can dramatically improve the efficiency of training and inference by routing each token to specialized expert modules that contain only a small fraction of...
[]
null
572
2103.16716
title_snapshot
v139/lezama21a
Run-Sort-ReRun: Escaping Batch Size Limitations in Sliced Wasserstein Generative Models
https://proceedings.mlr.press/v139/lezama21a.html
[ "Jose Lezama", "Wei Chen", "Qiang Qiu" ]
null
null
When training an implicit generative model, ideally one would like the generator to reproduce all the different modes and subtleties of the target distribution. Naturally, when comparing two empirical distributions, the larger the sample population, the more these statistical nuances can be captured. However, existing ...
[]
null
573
null
null
v139/li21a
PAGE: A Simple and Optimal Probabilistic Gradient Estimator for Nonconvex Optimization
https://proceedings.mlr.press/v139/li21a.html
[ "Zhize Li", "Hongyan Bao", "Xiangliang Zhang", "Peter Richtarik" ]
null
null
In this paper, we propose a novel stochastic gradient estimator—ProbAbilistic Gradient Estimator (PAGE)—for nonconvex optimization. PAGE is easy to implement as it is designed via a small adjustment to vanilla SGD: in each iteration, PAGE uses the vanilla minibatch SGD update with probability $p_t$ or reuses the previo...
[]
null
574
2008.10898
title_snapshot
v139/li21b
Tightening the Dependence on Horizon in the Sample Complexity of Q-Learning
https://proceedings.mlr.press/v139/li21b.html
[ "Gen Li", "Changxiao Cai", "Yuxin Chen", "Yuantao Gu", "Yuting Wei", "Yuejie Chi" ]
null
null
Q-learning, which seeks to learn the optimal Q-function of a Markov decision process (MDP) in a model-free fashion, lies at the heart of reinforcement learning. Focusing on the synchronous setting (such that independent samples for all state-action pairs are queried via a generative model in each iteration), substantia...
[]
null
575
null
null
v139/li21c
Winograd Algorithm for AdderNet
https://proceedings.mlr.press/v139/li21c.html
[ "Wenshuo Li", "Hanting Chen", "Mingqiang Huang", "Xinghao Chen", "Chunjing Xu", "Yunhe Wang" ]
null
null
Adder neural network (AdderNet) is a new kind of deep model that replaces the original massive multiplications in convolutions by additions while preserving the high performance. Since the hardware complexity of additions is much lower than that of multiplications, the overall energy consumption is thus reduced signifi...
[]
null
576
2105.05530
title_snapshot
v139/li21d
A Free Lunch From ANN: Towards Efficient, Accurate Spiking Neural Networks Calibration
https://proceedings.mlr.press/v139/li21d.html
[ "Yuhang Li", "Shikuang Deng", "Xin Dong", "Ruihao Gong", "Shi Gu" ]
null
null
Spiking Neural Network (SNN) has been recognized as one of the next generation of neural networks. Conventionally, SNN can be converted from a pre-trained ANN by only replacing the ReLU activation to spike activation while keeping the parameters intact. Perhaps surprisingly, in this work we show that a proper way to ca...
[]
null
577
2106.06984
title_snapshot
v139/li21e
Privacy-Preserving Feature Selection with Secure Multiparty Computation
https://proceedings.mlr.press/v139/li21e.html
[ "Xiling Li", "Rafael Dowsley", "Martine De Cock" ]
null
null
Existing work on privacy-preserving machine learning with Secure Multiparty Computation (MPC) is almost exclusively focused on model training and on inference with trained models, thereby overlooking the important data pre-processing stage. In this work, we propose the first MPC based protocol for private feature selec...
[]
null
578
2102.03517
title_snapshot
v139/li21f
Theory of Spectral Method for Union of Subspaces-Based Random Geometry Graph
https://proceedings.mlr.press/v139/li21f.html
[ "Gen Li", "Yuantao Gu" ]
null
null
Spectral method is a commonly used scheme to cluster data points lying close to Union of Subspaces, a task known as Subspace Clustering. The typical usage is to construct a Random Geometry Graph first and then apply spectral method to the graph to obtain clustering result. The latter step has been coined the name Spect...
[]
null
579
1907.10906
title_snapshot
v139/li21g
MURAL: Meta-Learning Uncertainty-Aware Rewards for Outcome-Driven Reinforcement Learning
https://proceedings.mlr.press/v139/li21g.html
[ "Kevin Li", "Abhishek Gupta", "Ashwin Reddy", "Vitchyr H Pong", "Aurick Zhou", "Justin Yu", "Sergey Levine" ]
null
null
Exploration in reinforcement learning is, in general, a challenging problem. A common technique to make learning easier is providing demonstrations from a human supervisor, but such demonstrations can be expensive and time-consuming to acquire. In this work, we study a more tractable class of reinforcement learning pro...
[]
null
580
2107.07184
title_snapshot
v139/li21h
Ditto: Fair and Robust Federated Learning Through Personalization
https://proceedings.mlr.press/v139/li21h.html
[ "Tian Li", "Shengyuan Hu", "Ahmad Beirami", "Virginia Smith" ]
null
null
Fairness and robustness are two important concerns for federated learning systems. In this work, we identify that robustness to data and model poisoning attacks and fairness, measured as the uniformity of performance across devices, are competing constraints in statistically heterogeneous networks. To address these con...
[]
null
581
2012.04221
title_snapshot
v139/li21i
Quantization Algorithms for Random Fourier Features
https://proceedings.mlr.press/v139/li21i.html
[ "Xiaoyun Li", "Ping Li" ]
null
null
The method of random projection (RP) is the standard technique for dimensionality reduction, approximate near neighbor search, compressed sensing, etc., which provides a simple and effective scheme for approximating pairwise inner products and Euclidean distances in massive data. Closely related to RP, the method of ra...
[]
null
582
2102.13079
title_snapshot
v139/li21j
Approximate Group Fairness for Clustering
https://proceedings.mlr.press/v139/li21j.html
[ "Bo Li", "Lijun Li", "Ankang Sun", "Chenhao Wang", "Yingfan Wang" ]
null
null
We incorporate group fairness into the algorithmic centroid clustering problem, where $k$ centers are to be located to serve $n$ agents distributed in a metric space. We refine the notion of proportional fairness proposed in [Chen et al., ICML 2019] as {\em core fairness}. A $k$-clustering is in the core if no coalitio...
[]
null
583
2203.17146
title_snapshot
v139/li21k
Sharper Generalization Bounds for Clustering
https://proceedings.mlr.press/v139/li21k.html
[ "Shaojie Li", "Yong Liu" ]
null
null
Existing generalization analysis of clustering mainly focuses on specific instantiations, such as (kernel) $k$-means, and a unified framework for studying clustering performance is still lacking. Besides, the existing excess clustering risk bounds are mostly of order $\mathcal{O}(K/\sqrt{n})$ provided that the underlyi...
[]
null
584
null
null
v139/li21l
Provably End-to-end Label-noise Learning without Anchor Points
https://proceedings.mlr.press/v139/li21l.html
[ "Xuefeng Li", "Tongliang Liu", "Bo Han", "Gang Niu", "Masashi Sugiyama" ]
null
null
In label-noise learning, the transition matrix plays a key role in building statistically consistent classifiers. Existing consistent estimators for the transition matrix have been developed by exploiting anchor points. However, the anchor-point assumption is not always satisfied in real scenarios. In this paper, we pr...
[]
null
585
2102.02400
title_snapshot
v139/li21m
A Novel Method to Solve Neural Knapsack Problems
https://proceedings.mlr.press/v139/li21m.html
[ "Duanshun Li", "Jing Liu", "Dongeun Lee", "Ali Seyedmazloom", "Giridhar Kaushik", "Kookjin Lee", "Noseong Park" ]
null
null
0-1 knapsack is of fundamental importance across many fields. In this paper, we present a game-theoretic method to solve 0-1 knapsack problems (KPs) where the number of items (products) is large and the values of items are not predetermined but decided by an external value assignment function (e.g., a neural network in...
[]
null
586
null
null
v139/li21n
Mixed Cross Entropy Loss for Neural Machine Translation
https://proceedings.mlr.press/v139/li21n.html
[ "Haoran Li", "Wei Lu" ]
null
null
In neural machine translation, Cross Entropy loss (CE) is the standard loss function in two training methods of auto-regressive models, i.e., teacher forcing and scheduled sampling. In this paper, we propose mixed Cross Entropy loss (mixed CE) as a substitute for CE in both training approaches. In teacher forcing, the ...
[]
null
587
2106.15880
title_snapshot
v139/li21o
Training Graph Neural Networks with 1000 Layers
https://proceedings.mlr.press/v139/li21o.html
[ "Guohao Li", "Matthias Müller", "Bernard Ghanem", "Vladlen Koltun" ]
null
null
Deep graph neural networks (GNNs) have achieved excellent results on various tasks on increasingly large graph datasets with millions of nodes and edges. However, memory complexity has become a major obstacle when training deep GNNs for practical applications due to the immense number of nodes, edges, and intermediate ...
[]
null
588
2106.07476
title_snapshot
v139/li21p
Active Feature Acquisition with Generative Surrogate Models
https://proceedings.mlr.press/v139/li21p.html
[ "Yang Li", "Junier Oliva" ]
null
null
Many real-world situations allow for the acquisition of additional relevant information when making an assessment with limited or uncertain data. However, traditional ML approaches either require all features to be acquired beforehand or regard part of them as missing data that cannot be acquired. In this work, we cons...
[]
null
589
2010.02433
title_snapshot
v139/li21q
Partially Observed Exchangeable Modeling
https://proceedings.mlr.press/v139/li21q.html
[ "Yang Li", "Junier Oliva" ]
null
null
Modeling dependencies among features is fundamental for many machine learning tasks. Although there are often multiple related instances that may be leveraged to inform conditional dependencies, typical approaches only model conditional dependencies over individual instances. In this work, we propose a novel framework,...
[]
null
590
2102.06083
title_snapshot
v139/li21r
Testing DNN-based Autonomous Driving Systems under Critical Environmental Conditions
https://proceedings.mlr.press/v139/li21r.html
[ "Zhong Li", "Minxue Pan", "Tian Zhang", "Xuandong Li" ]
null
null
Due to the increasing usage of Deep Neural Network (DNN) based autonomous driving systems (ADS) where erroneous or unexpected behaviours can lead to catastrophic accidents, testing such systems is of growing importance. Existing approaches often just focus on finding erroneous behaviours and have not thoroughly studied...
[]
null
591
null
null
v139/li21s
The Symmetry between Arms and Knapsacks: A Primal-Dual Approach for Bandits with Knapsacks
https://proceedings.mlr.press/v139/li21s.html
[ "Xiaocheng Li", "Chunlin Sun", "Yinyu Ye" ]
null
null
In this paper, we study the bandits with knapsacks (BwK) problem and develop a primal-dual based algorithm that achieves a problem-dependent logarithmic regret bound. The BwK problem extends the multi-arm bandit (MAB) problem to model the resource consumption, and the existing BwK literature has been mainly focused on ...
[]
null
592
2102.06385
title_snapshot
v139/li21t
Distributionally Robust Optimization with Markovian Data
https://proceedings.mlr.press/v139/li21t.html
[ "Mengmeng Li", "Tobias Sutter", "Daniel Kuhn" ]
null
null
We study a stochastic program where the probability distribution of the uncertain problem parameters is unknown and only indirectly observed via finitely many correlated samples generated by an unknown Markov chain with $d$ states. We propose a data-driven distributionally robust optimization model to estimate the prob...
[]
null
593
2106.06741
title_snapshot
v139/li21u
Communication-Efficient Distributed SVD via Local Power Iterations
https://proceedings.mlr.press/v139/li21u.html
[ "Xiang Li", "Shusen Wang", "Kun Chen", "Zhihua Zhang" ]
null
null
We study distributed computing of the truncated singular value decomposition (SVD). We develop an algorithm that we call \texttt{LocalPower} for improving communication efficiency. Specifically, we uniformly partition the dataset among $m$ nodes and alternate between multiple (precisely $p$) local power iterations and ...
[]
null
594
2002.08014
title_snapshot
v139/li21v
FILTRA: Rethinking Steerable CNN by Filter Transform
https://proceedings.mlr.press/v139/li21v.html
[ "Bo Li", "Qili Wang", "Gim Hee Lee" ]
null
null
Steerable CNN imposes the prior knowledge of transformation invariance or equivariance in the network architecture to enhance the the network robustness on geometry transformation of data and reduce overfitting. It has been an intuitive and widely used technique to construct a steerable filter by augmenting a filter wi...
[]
null
595
2105.11636
title_snapshot
v139/li21w
Online Unrelated Machine Load Balancing with Predictions Revisited
https://proceedings.mlr.press/v139/li21w.html
[ "Shi Li", "Jiayi Xian" ]
null
null
We study the online load balancing problem with machine learned predictions, and give results that improve upon and extend those in a recent paper by Lattanzi et al. (2020). First, we design deterministic and randomized online rounding algorithms for the problem in the unrelated machine setting, with $O(\frac{\log m}{\...
[]
null
596
null
null
v139/li21x
Asymptotic Normality and Confidence Intervals for Prediction Risk of the Min-Norm Least Squares Estimator
https://proceedings.mlr.press/v139/li21x.html
[ "Zeng Li", "Chuanlong Xie", "Qinwen Wang" ]
null
null
This paper quantifies the uncertainty of prediction risk for the min-norm least squares estimator in high-dimensional linear regression models. We establish the asymptotic normality of prediction risk when both the sample size and the number of features tend to infinity. Based on the newly established central limit the...
[]
null
597
null
null
v139/li21y
TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models
https://proceedings.mlr.press/v139/li21y.html
[ "Zhuohan Li", "Siyuan Zhuang", "Shiyuan Guo", "Danyang Zhuo", "Hao Zhang", "Dawn Song", "Ion Stoica" ]
null
null
Model parallelism has become a necessity for training modern large-scale deep language models. In this work, we identify a new and orthogonal dimension from existing model parallel approaches: it is possible to perform pipeline parallelism within a single training sequence for Transformer-based language models thanks t...
[]
null
598
2102.07988
title_snapshot
v139/li21z
A Second look at Exponential and Cosine Step Sizes: Simplicity, Adaptivity, and Performance
https://proceedings.mlr.press/v139/li21z.html
[ "Xiaoyu Li", "Zhenxun Zhuang", "Francesco Orabona" ]
null
null
Stochastic Gradient Descent (SGD) is a popular tool in training large-scale machine learning models. Its performance, however, is highly variable, depending crucially on the choice of the step sizes. Accordingly, a variety of strategies for tuning the step sizes have been proposed, ranging from coordinate-wise approach...
[]
null
599
2002.05273
title_snapshot
v139/liang21a
Towards Understanding and Mitigating Social Biases in Language Models
https://proceedings.mlr.press/v139/liang21a.html
[ "Paul Pu Liang", "Chiyu Wu", "Louis-Philippe Morency", "Ruslan Salakhutdinov" ]
null
null
As machine learning methods are deployed in real-world settings such as healthcare, legal systems, and social science, it is crucial to recognize how they shape social biases and stereotypes in these sensitive decision-making processes. Among such real-world deployments are large-scale pretrained language models (LMs) ...
[]
null
600
2106.13219
title_snapshot