Search Results for author:
Found 47 papers, 30 papers with code
Overparameterization hurts worst-group accuracy with spurious correlations
Increasing model capacity well beyond the point of zero training error has been observed to improve average test accuracy.
Scaling Laws for Data Filtering -- Data Curation cannot be Compute Agnostic
Vision-language models (VLMs) are trained for thousands of GPU hours on carefully curated web datasets.
Predicting the Performance of Foundation Models via Agreement-on-the-Line
Recently, it was shown that ensembles of neural networks observe the phenomena ``agreement-on-the-line'', which can be leveraged to reliably predict OOD performance without labels.
Jailbreaking is Best Solved by Definition
The rise of "jailbreak" attacks on language models has led to a flurry of defenses aimed at preventing the output of undesirable responses.
Repetition Improves Language Model Embeddings
In this work, we address an architectural limitation of autoregressive models: token embeddings cannot contain information from tokens that appear later in the input.
AutoFT: Learning an Objective for Robust Fine-Tuning
Given a task, AutoFT searches for a fine-tuning procedure that enhances out-of-distribution (OOD) generalization.
Complementary Benefits of Contrastive Learning and Self-Training Under Distribution Shift
Self-training and contrastive learning have emerged as leading techniques for incorporating unlabeled data, both under distribution shift (unsupervised domain adaptation) and when it is absent (semi-supervised learning).
Multitask Learning Can Improve Worst-Group Outcomes
We primarily consider the standard setting of fine-tuning a pre-trained model, where, following recent work \citep{gururangan2020don, dery2023aang}, we multitask the end task with the pre-training objective constructed from the end task data itself.
Reliable Test-Time Adaptation via Agreement-on-the-Line
In this work, we make a notable and surprising observation that TTAed models strongly show the agreement-on-the-line phenomenon (Baek et al., 2022) across a wide range of distribution shifts.
Understanding Catastrophic Forgetting in Language Models via Implicit Inference
We lack a systematic understanding of the effects of fine-tuning (via methods such as instruction-tuning or reinforcement learning from human feedback), particularly on tasks outside the narrow fine-tuning distribution.
Contextual Reliability: When Different Features Matter in Different Contexts
We formalize a new setting called contextual reliability which accounts for the fact that the "right" features to use may vary depending on the context.
T-MARS: Improving Visual Representations by Circumventing Text Feature Learning
However, naively removing all such data could also be wasteful, as it throws away images that contain visual features (in addition to overlapping text).
ALP: Action-Aware Embodied Learning for Perception
In addition, we show that by training on actively collected data more relevant to the environment and task, our method generalizes more robustly to downstream tasks compared to models pre-trained on fixed datasets such as ImageNet.
Automatically Auditing Large Language Models via Discrete Optimization
Auditing large language models for unexpected behaviors is critical to preempt catastrophic deployments, yet remains challenging.
Bitrate-Constrained DRO: Beyond Worst Case Robustness To Unknown Group Shifts
Some robust training algorithms (e. g., Group DRO) specialize to group shifts and require group information on all training points.
Learning Representations that Enable Generalization in Assistive Tasks
We advocate that generalization to such OOD policies benefits from (1) learning a good latent representation for human policies that test-time humans can accurately be mapped to, and (2) making that representation adaptable with test-time interaction data, instead of relying on it to perfectly capture the space of human policies based on the simulated population only.
Finetune like you pretrain: Improved finetuning of zero-shot vision models
In total, these benchmarks establish contrastive finetuning as a simple, intuitive, and state-of-the-art approach for supervised finetuning of image-text models like CLIP.
Using Language to Extend to Unseen Domains
It is expensive to collect training data for every possible domain that a vision model may encounter when deployed.
Test-Time Adaptation via Conjugate Pseudo-labels
In this paper, we start by presenting a surprising phenomenon: if we attempt to meta-learn the best possible TTA loss over a wide class of functions, then we recover a function that is remarkably similar to (a temperature-scaled version of) the softmax-entropy employed by TENT.
Calibrated ensembles can mitigate accuracy tradeoffs under distribution shift
We often see undesirable tradeoffs in robust machine learning where out-of-distribution (OOD) accuracy is at odds with in-distribution (ID) accuracy: a robust classifier obtained via specialized techniques such as removing spurious features often has better OOD but worse ID accuracy compared to a standard classifier trained via ERM.
Agreement-on-the-Line: Predicting the Performance of Neural Networks under Distribution Shift
In this paper, we show a similar but surprising phenomenon also holds for the agreement between pairs of neural network classifiers: whenever accuracy-on-the-line holds, we observe that the OOD agreement between the predictions of any two pairs of neural networks (with potentially different architectures) also observes a strong linear correlation with their ID agreement.
Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution
However, in this paper, we find that fine-tuning can achieve worse accuracy than linear probing out-of-distribution (OOD) when the pretrained features are good and the distribution shift is large.
An Explanation of In-context Learning as Implicit Bayesian Inference
At test time, in-context learning occurs when the LM also infers a shared latent concept between examples in a prompt.
Fine-Tuning Distorts Pretrained Features and Underperforms Out-of-Distribution
It is well known that fine-tuning leads to better accuracy in-distribution (ID).
Calibrated ensembles - a simple way to mitigate ID-OOD accuracy tradeoffs
We often see undesirable tradeoffs in robust machine learning where out-of-distribution (OOD) accuracy is at odds with in-distribution (ID) accuracy.
On the Opportunities and Risks of Foundation Models
AI is undergoing a paradigm shift with the rise of models (e. g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks.
Just Train Twice: Improving Group Robustness without Training Group Information
Standard training via empirical risk minimization (ERM) can produce models that achieve high accuracy on average but low accuracy on certain groups, especially in the presence of spurious correlations between the input and label.
Ranked #1 on
Out-of-Distribution Generalization
on ImageNet-W
Accuracy on the Line: On the Strong Correlation Between Out-of-Distribution and In-Distribution Generalization
For machine learning systems to be reliable, we must understand their performance in unseen, out-of-distribution environments.
Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming
In this work, we propose a first-order dual SDP algorithm that (1) requires memory only linear in the total number of network activations, (2) only requires a fixed number of forward/backward passes through the network per iteration.
Decoupling Exploration and Exploitation for Meta-Reinforcement Learning without Sacrifices
Learning a new task often requires both exploring to gather task-relevant information and exploiting this information to solve the task.
The Pitfalls of Simplicity Bias in Neural Networks
Furthermore, previous settings that use SB to theoretically justify why neural networks generalize well do not simultaneously capture the non-robustness of neural networks---a widely observed phenomenon in practice [Goodfellow et al. 2014, Jo and Bengio 2017].
Explore then Execute: Adapting without Rewards via Factorized Meta-Reinforcement Learning
In principle, meta-reinforcement learning approaches can exploit this shared structure, but in practice, they fail to adapt to new environments when adaptation requires targeted exploration (e. g., exploring the cabinets to find ingredients in a new kitchen).
An Investigation of Why Overparameterization Exacerbates Spurious Correlations
We study why overparameterization -- increasing model size well beyond the point of zero training error -- can hurt test error on minority groups despite improving average test error when there are spurious correlations in the data.
Robust Encodings: A Framework for Combating Adversarial Typos
We instantiate RobEn to defend against a large family of adversarial typos.
DROCC: Deep Robust One-Class Classification
Classical approaches for one-class problems such as one-class SVM and isolation forest require careful feature engineering when applied to structured domains like images.
Ranked #3 on
Anomaly Detection
on UEA time-series datasets
Understanding and Mitigating the Tradeoff Between Robustness and Accuracy
In this work, we precisely characterize the effect of augmentation on the standard error in linear regression when the optimal linear predictor has zero standard and robust error.
Certified Robustness to Adversarial Word Substitutions
We train the first models that are provably robust to all word substitutions in this family.
Adversarial Training Can Hurt Generalization
While adversarial training can improve robust accuracy (against an adversary), it sometimes hurts standard accuracy (when there is no adversary).
Maximum Weighted Loss Discrepancy
To capture this inequality, we introduce and study a notion we call maximum weighted loss discrepancy (MWLD), the maximum (weighted) difference between the loss of a group and the loss of the population.
Unlabeled Data Improves Adversarial Robustness
We demonstrate, theoretically and empirically, that adversarial robustness can significantly benefit from semisupervised learning.
Semidefinite relaxations for certifying robustness to adversarial examples
One promise of ending the arms race is developing certified defenses, ones which are provably robust against all attackers in some family.
Certified Defenses against Adversarial Examples
While neural networks have achieved high accuracy on standard image classification benchmarks, their accuracy drops to nearly zero in the presence of small adversarial perturbations to test inputs.
Estimating the unseen from multiple populations
We generalize this extrapolation and related unseen estimation problems to the multiple population setting, where population $j$ has an unknown distribution $D_j$ from which we observe $n_j$ samples.
Learning Mixture of Gaussians with Streaming Data
However, by using a streaming version of the classical (soft-thresholding-based) EM method that exploits the Gaussian distribution explicitly, we show that for a mixture of two Gaussians the true means can be estimated consistently, with estimation error decreasing at nearly optimal rate, and tending to $0$ for $N\rightarrow \infty$.
Estimation from Indirect Supervision with Linear Moments
In structured prediction problems where we have indirect supervision of the output, maximum marginal likelihood faces two computational obstacles: non-convexity of the objective and intractability of even a single gradient computation.
Probabilistic Dependency Networks for Prediction and Diagnostics
Research in transportation frequently involve modelling and predicting attributes of events that occur at regular intervals.
A Reinforcement Learning Approach to Online Learning of Decision Trees
Online decision tree learning algorithms typically examine all features of a new data point to update model parameters.
