Author name cluster

John Canny

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

8 papers

1 author row

CLeaR Conference 2022 Conference Paper

Learning Causal Overhypotheses through Exploration in Children and Computational Models

Eliza Kosoy
Adrian Liu
Jasmine L Collins
David Chan
Jessica B Hamrick
Nan Rosemary Ke
Sandy Huang
Bryanna Kaufmann

Despite recent progress in reinforcement learning (RL), RL algorithms for exploration still remain an active area of research. Existing methods often focus on state-based metrics, which do not con-sider the underlying causal structures of the environment, and while recent research has begun to explore RL environments for causal learning, these environments primarily leverage causal information through causal inference or induction rather than exploration. In contrast, human children—some of the most proficient explorers—have been shown to use causal information to great benefit. In this work, we introduce a novel RL environment designed with a controllable causal structure, which allows us to evaluate exploration strategies used by both agents and children in a unified environment. In addition, through experimentation on both computation models and children, we demonstrate that there are significant differences between information-gain optimal RL exploration in causal environments and the exploration of children in the same environments. We leverage this new insight to lay the groundwork for future research into efficient exploration and disambiguation of causal structures for RL algorithms.

PDF Details

NeurIPS Conference 2021 Conference Paper

Compressive Visual Representations

Kuang-Huei Lee
Anurag Arnab
Sergio Guadarrama
John Canny
Ian Fischer

Learning effective visual representations that generalize well without human supervision is a fundamental problem in order to apply Machine Learning to a wide variety of tasks. Recently, two families of self-supervised methods, contrastive learning and latent bootstrapping, exemplified by SimCLR and BYOL respectively, have made significant progress. In this work, we hypothesize that adding explicit information compression to these algorithms yields better and more robust representations. We verify this by developing SimCLR and BYOL formulations compatible with the Conditional Entropy Bottleneck (CEB) objective, allowing us to both measure and control the amount of compression in the learned representation, and observe their impact on downstream tasks. Furthermore, we explore the relationship between Lipschitz continuity and compression, showing a tractable lower bound on the Lipschitz constant of the encoders we learn. As Lipschitz continuity is closely related to robustness, this provides a new explanation for why compressed models are more robust. Our experiments confirm that adding compression to SimCLR and BYOL significantly improves linear evaluation accuracies and model robustness across a wide range of domain shifts. In particular, the compressed version of BYOL achieves 76. 0% Top-1 linear evaluation accuracy on ImageNet with ResNet-50, and 78. 8% with ResNet-50 2x.

PDF Details

NeurIPS Conference 2020 Conference Paper

Predictive Information Accelerates Learning in RL

Kuang-Huei Lee
Ian Fischer
Anthony Liu
Yijie Guo
Honglak Lee
John Canny
Sergio Guadarrama

The Predictive Information is the mutual information between the past and the future, I(X past; X future). We hypothesize that capturing the predictive information is useful in RL, since the ability to model what will happen next is necessary for success on many tasks. To test our hypothesis, we train Soft Actor-Critic (SAC) agents from pixels with an auxiliary task that learns a compressed representation of the predictive information of the RL environment dynamics using a contrastive version of the Conditional Entropy Bottleneck (CEB) objective. We refer to these as Predictive Information SAC (PI-SAC) agents. We show that PI-SAC agents can substantially improve sample efficiency over challenging baselines on tasks from the DM Control suite of continuous control environments. We evaluate PI-SAC agents by comparing against uncompressed PI-SAC agents, other compressed and uncompressed agents, and SAC agents directly trained from pixels. Our implementation is given on GitHub.

PDF Details

NeurIPS Conference 2019 Conference Paper

Evaluating Protein Transfer Learning with TAPE

Roshan Rao
Nicholas Bhattacharya
Neil Thomas
Yan Duan
Peter Chen
John Canny
Pieter Abbeel
Yun Song

Protein modeling is an increasingly popular area of machine learning research. Semi-supervised learning has emerged as an important paradigm in protein modeling due to the high cost of acquiring supervised protein labels, but the current literature is fragmented when it comes to datasets and standardized evaluation techniques. To facilitate progress in this field, we introduce the Tasks Assessing Protein Embeddings (TAPE), a set of five biologically relevant semi-supervised learning tasks spread across different domains of protein biology. We curate tasks into specific training, validation, and test splits to ensure that each task tests biologically relevant generalization that transfers to real-life scenarios. We benchmark a range of approaches to semi-supervised protein representation learning, which span recent work as well as canonical sequence learning techniques. We find that self-supervised pretraining is helpful for almost all models on all tasks, more than doubling performance in some cases. Despite this increase, in several cases features learned by self-supervised pretraining still lag behind features extracted by state-of-the-art non-neural techniques. This gap in performance suggests a huge opportunity for innovative architecture design and improved modeling paradigms that better capture the signal in biological sequences. TAPE will help the machine learning community focus effort on scientifically relevant problems. Toward this end, all data and code used to run these experiments is available at https: //github. com/songlab-cal/tape

PDF Details

IJCAI Conference 2018 Conference Paper

An Efficient Minibatch Acceptance Test for Metropolis-Hastings

Daniel Seita
Xinlei Pan
Haoyu Chen
John Canny

We present a novel Metropolis-Hastings method for large datasets that uses small expected-size mini-batches of data. Previous work on reducing the cost of Metropolis-Hastings tests yields only constant factor reductions versus using the full dataset for each sample. Here we present a method that can be tuned to provide arbitrarily small batch sizes, by adjusting either proposal step size or temperature. Our test uses the noise-tolerant Barker acceptance test with a novel additive correction variable. The resulting test has similar cost to a normal SGD update. Our experiments demonstrate several order-of-magnitude speedups over previous work.

PDF Details

NeurIPS Conference 2009 Conference Paper

Factor Modeling for Advertisement Targeting

Ye Chen
Michael Kapralov
John Canny
Dmitry Pavlov

We adapt a probabilistic latent variable model, namely GaP (Gamma-Poisson), to ad targeting in the contexts of sponsored search (SS) and behaviorally targeted (BT) display advertising. We also approach the important problem of ad positional bias by formulating a one-latent-dimension GaP factorization. Learning from click-through data is intrinsically large scale, even more so for ads. We scale up the algorithm to terabytes of real-world SS and BT data that contains hundreds of millions of users and hundreds of thousands of features, by leveraging the scalability characteristics of the algorithm and the inherent structure of the problem including data sparsity and locality. Specifically, we demonstrate two somewhat orthogonal philosophies of scaling algorithms to large-scale problems, through the SS and BT implementations, respectively. Finally, we report the experimental results using Yahoos vast datasets, and show that our approach substantially outperform the state-of-the-art methods in prediction accuracy. For BT in particular, the ROC area achieved by GaP is exceeding 0. 95, while one prior approach using Poisson regression yielded 0. 83. For computational performance, we compare a single-node sparse implementation with a parallel implementation using Hadoop MapReduce, the results are counterintuitive yet quite interesting. We therefore provide insights into the underlying principles of large-scale learning.

PDF Details

AIJ Journal 1988 Journal Article

Constructing roadmaps of semi-algebraic sets I: Completeness

John Canny

This paper describes preliminary work on an algorithm for planning collision-free motions for a robot manipulator in the presence of obstacles. The physical obstacles lead to forbidden regions in the robots configuration space, and for collision-free motion we need paths through configuration space which avoid these regions. Our method is to construct a certain one-dimensional subset or “roadmap” of the space of allowable configurations. If S denotes the set of allowable configurations, the roadmap has the property that any connected component of S contains a single connected component of the roadmap. It is also possible, starting from an arbitrary point p ∈ S to rapidly construct a path from p to a point on the roadmap. Thus given any two points in S we can rapidly determine whether they lie in the same connected component of S, and if they do, we can return a candidate path between them. We do not give a complete description of the algorithm here, but we define the roadmap geometrically, and verify that it has the necessary connectivity.

Details DOI

AAAI Conference 1983 Conference Paper

A Variational Approach to Edge Detection

John Canny

The problem of detecting intensity changes in images is canonical in vision. Edge detection operators are typically designed to optimally estimate first or second derivative over some (usually small) support. Other criteria such as output signal to noise ratio or bandwidth have also been argued for. This paper describes an attempt to formulate set of edge detection criteria that capture as directly as possible the desirable properties of the detector. Variational techniques are used to find 2 solution over the space of all possible functions. The first criterion is that the detector have low probability of error i.e. failing to mark edges or falsely marking non-edges. The second is that the marked points should be as close as possible to the centre of the true edge. The third criterion is that there should be low probability of more than one response to a single edge. The third criterion is claimed to be new, and it became necessary when an operator designed using the first two criteria was found to have excessive multiple responses. The edge model that will be considered here is 2 one-dlmensional step edge in white Gaussian noise although the same technique has been applied to an extended impulse or ridge profile. The result is a one dimensional operator that approximates the first derivative of a Gaussian. Its extension to two dimensions is also discussed.

PDF Details