4.3k citations · 10.2k across the 23 of their papers we have counts for
26 papers
Seasoning Model Soups for Robustness to Adversarial and Natural Distribution Shifts
Francesco Croce, Sylvestre-Alvise Rebuffi, Evan Shelhamer +1
Adversarial training is widely used to make classifiers robust to a specific threat or adversary, such as -norm bounded perturbations of a given -norm. However, existing…
Where Should I Spend My FLOPS? Efficiency Evaluations of Visual Pre-training Methods
Skanda Koppula, Yazhe Li, Evan Shelhamer +5
Self-supervised methods have achieved remarkable success in transfer learning, often achieving the same or better accuracy than supervised pre-training. Most prior work has done so…
Back to the Source: Diffusion-Driven Test-Time Adaptation
Jin Gao, Jialing Zhang, Xihui Liu +3
Test-time adaptation harnesses test inputs to improve the accuracy of a model trained on source data when tested on shifted target data. Existing methods update the source model by…
Object discovery and representation networks
Olivier J. Hénaff, Skanda Koppula, Evan Shelhamer +5
The promise of self-supervised learning (SSL) is to leverage large amounts of unlabeled data to solve complex tasks. While there has been excellent progress with simple, image-leve…
Evaluating the Adversarial Robustness of Adaptive Test-time Defenses
Francesco Croce, Sven Gowal, Thomas Brunner +3
Adaptive defenses, which optimize at test time, promise to improve adversarial robustness. We categorize such adaptive test-time defenses, explain their potential benefits and draw…
HiP: Hierarchical Perceiver
Joao Carreira, Skanda Koppula, Daniel Zoran +10
General perception systems such as Perceivers can process arbitrary modalities in any combination and are able to handle up to a few hundred thousand inputs. They achieve this gene…
On-target Adaptation
Dequan Wang, Shaoteng Liu, Sayna Ebrahimi +2
Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data…
Perceiver IO: A General Architecture for Structured Inputs & Outputs
Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac +12
A central goal of machine learning is the development of systems that can solve many problems in as many data domains as possible. Current architectures, however, cannot be applied…
Fighting Gradients with Gradients: Dynamic Defenses against Adversarial Attacks
Dequan Wang, An Ju, Evan Shelhamer +2
Adversarial attacks optimize against models to defeat defenses. Existing defenses are static, and stay the same once trained, even while attacks change. We argue that models should…
Anytime Dense Prediction with Confidence Adaptivity
Zhuang Liu, Zhiqiu Xu, Hung-Ju Wang +2
Anytime inference requires a model to make a progression of predictions which might be halted at any time. Prior research on anytime visual recognition has mostly focused on image…
It Is Likely That Your Loss Should be a Likelihood
Mark Hamilton, Evan Shelhamer, William T. Freeman
Many common loss functions such as mean-squared-error, cross-entropy, and reconstruction loss are unnecessarily rigid. Under a probabilistic interpretation, these common losses cor…
Tent: Fully Test-time Adaptation by Entropy Minimization
Dequan Wang, Evan Shelhamer, Shaoteng Liu +2
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own paramet…
Exploring Simple and Transferable Recognition-Aware Image Processing
Zhuang Liu, Hung-Ju Wang, Tinghui Zhou +4
Recent progress in image recognition has stimulated the deployment of vision systems at an unprecedented scale. As a result, visual data are now often consumed not only by humans b…
Dynamic Scale Inference by Entropy Minimization
Dequan Wang, Evan Shelhamer, Bruno Olshausen +1
Given the variety of the visual world there is not one true scale for recognition: objects may appear at drastically different sizes across the visual field. Rather than enumerate…
Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields
Evan Shelhamer, Dequan Wang, Trevor Darrell
The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end…
Infinite Mixture Prototypes for Few-Shot Learning
Kelsey R. Allen, Evan Shelhamer, Hanul Shin +1
We propose infinite mixture prototypes to adaptively represent both simple and complex data distributions for few-shot learning. Our infinite mixture prototypes represent each clas…
Few-Shot Segmentation Propagation with Guided Networks
Kate Rakelly, Evan Shelhamer, Trevor Darrell +2
Learning-based methods for visual segmentation have made progress on particular types of segmentation tasks, but are limited by the necessary supervision, the narrow definitions of…
Zero-Shot Visual Imitation
Deepak Pathak, Parsa Mahmoudieh, Guanghao Luo +7
The current dominant paradigm for imitation learning relies on strong supervision of expert actions to learn both 'what' and 'how' to imitate. We pursue an alternative paradigm whe…
Loss is its own Reward: Self-Supervision for Reinforcement Learning
Evan Shelhamer, Parsa Mahmoudieh, Max Argus +1
Reinforcement learning optimizes policies for expected cumulative reward. Need the supervision be so narrow? Reward is delayed and sparse for many tasks, making it a difficult and…
Clockwork Convnets for Video Semantic Segmentation
Evan Shelhamer, Kate Rakelly, Judy Hoffman +1
Recent years have seen tremendous progress in still-image segmentation; however the naïve application of these state-of-the-art algorithms to every video frame requires considerabl…
Fully Convolutional Networks for Semantic Segmentation
Evan Shelhamer, Jonathan Long, Trevor Darrell
Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, impro…
Fine-grained pose prediction, normalization, and recognition
Ning Zhang, Evan Shelhamer, Yang Gao +1
Pose variation and subtle differences in appearance are key challenges to fine-grained classification. While deep networks have markedly improved general recognition, many approach…
Fully Convolutional Multi-Class Multiple Instance Learning
Deepak Pathak, Evan Shelhamer, Jonathan Long +1
Multiple instance learning (MIL) can reduce the need for costly annotation in tasks such as semantic segmentation by weakening the required degree of supervision. We propose a nove…
Fully Convolutional Networks for Semantic Segmentation
Jonathan Long, Evan Shelhamer, Trevor Darrell
Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, excee…