arXiv subjects

Percy Liang

Publications and source records attributed to Percy Liang.

At least 181 records · Page 10Linked to original sources

arXiv2021

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. Prior approaches that achieve high worst-group accuracy, like group distributionally robust optimization (group DRO) require expensive group annotations for each training point, whereas approaches that do not use such group annotations typically achieve unsatisfactory worst-group accuracy. In this paper, we propose a simple two-stage approach, JTT, that first trains a standard ERM model for several epochs, and then trains a second model that upweights the training examples that the first model misclassified. Intuitively, this upweights examples from groups on which standard ERM models perform poorly, leading to improved worst-group performance. Averaged over four image classification and natural language processing tasks with spurious correlations, JTT closes 75% of the gap in worst-group accuracy between standard ERM and group DRO, while only requiring group annotations on a small validation set in order to tune hyperparameters.

cs.LG↗

arXiv2021

Conditional probing: measuring usable information beyond a baseline

Probing experiments investigate the extent to which neural representations make properties -- like part-of-speech -- predictable. One suggests that a representation encodes a property if probing that representation produces higher accuracy than probing a baseline representation like non-contextual word embeddings. Instead of using baselines as a point of comparison, we're interested in measuring information that is contained in the representation but not in the baseline. For example, current methods can detect when a representation is more useful than the word identity (a baseline) for predicting part-of-speech; however, they cannot detect when the representation is predictive of just the aspects of part-of-speech not explainable by the word identity. In this work, we extend a theory of usable information called $\mathcal{V}$-information and propose conditional probing, which explicitly conditions on the information in the baseline. In a case study, we find that after conditioning on non-contextual word embeddings, properties like part-of-speech are accessible at deeper layers of a network than previously thought.

cs.CL↗

arXiv2021

No True State-of-the-Art? OOD Detection Methods are Inconsistent across Datasets

Out-of-distribution detection is an important component of reliable ML systems. Prior literature has proposed various methods (e.g., MSP (Hendrycks & Gimpel, 2017), ODIN (Liang et al., 2018), Mahalanobis (Lee et al., 2018)), claiming they are state-of-the-art by showing they outperform previous methods on a selected set of in-distribution (ID) and out-of-distribution (OOD) datasets. In this work, we show that none of these methods are inherently better at OOD detection than others on a standardized set of 16 (ID, OOD) pairs. We give possible explanations for these inconsistencies with simple toy datasets where whether one method outperforms another depends on the structure of the ID and OOD datasets in question. Finally, we show that a method outperforming another on a certain (ID, OOD) pair may not do so in a low-data regime. In the low-data regime, we propose a distance-based method, Pairwise OOD detection (POD), which is based on Siamese networks and improves over Mahalanobis by sidestepping the expensive covariance estimation step. Our results suggest that the OOD detection problem may be too broad, and we should consider more specific structures for leverage.

cs.LG↗

arXiv2021

WILDS: A Benchmark of in-the-Wild Distribution Shifts

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 community today. To address this gap, we present WILDS, a curated benchmark of 10 datasets reflecting a diverse range of distribution shifts that naturally arise in real-world applications, such as shifts across hospitals for tumor identification; across camera traps for wildlife monitoring; and across time and location in satellite imaging and poverty mapping. On each dataset, we show that standard training yields substantially lower out-of-distribution than in-distribution performance. This gap remains even with models trained by existing methods for tackling distribution shifts, underscoring the need for new methods for training models that are more robust to the types of distribution shifts that arise in practice. To facilitate method development, we provide an open-source package that automates dataset loading, contains default model architectures and hyperparameters, and standardizes evaluations. Code and leaderboards are available at https://wilds.stanford.edu.

cs.LG↗

arXiv2021

Codified audio language modeling learns useful representations for music information retrieval

We demonstrate that language models pre-trained on codified (discretely-encoded) music audio learn representations that are useful for downstream MIR tasks. Specifically, we explore representations from Jukebox (Dhariwal et al. 2020): a music generation system containing a language model trained on codified audio from 1M songs. To determine if Jukebox's representations contain useful information for MIR, we use them as input features to train shallow models on several MIR tasks. Relative to representations from conventional MIR models which are pre-trained on tagging, we find that using representations from Jukebox as input features yields 30% stronger performance on average across four MIR tasks: tagging, genre classification, emotion recognition, and key detection. For key detection, we observe that representations from Jukebox are considerably stronger than those from models pre-trained on tagging, suggesting that pre-training via codified audio language modeling may address blind spots in conventional approaches. We interpret the strength of Jukebox's representations as evidence that modeling audio instead of tags provides richer representations for MIR.

cs.SD↗

arXiv2021

Break-It-Fix-It: Unsupervised Learning for Program Repair

We consider repair tasks: given a critic (e.g., compiler) that assesses the quality of an input, the goal is to train a fixer that converts a bad example (e.g., code with syntax errors) into a good one (e.g., code with no syntax errors). Existing works create training data consisting of (bad, good) pairs by corrupting good examples using heuristics (e.g., dropping tokens). However, fixers trained on this synthetically-generated data do not extrapolate well to the real distribution of bad inputs. To bridge this gap, we propose a new training approach, Break-It-Fix-It (BIFI), which has two key ideas: (i) we use the critic to check a fixer's output on real bad inputs and add good (fixed) outputs to the training data, and (ii) we train a breaker to generate realistic bad code from good code. Based on these ideas, we iteratively update the breaker and the fixer while using them in conjunction to generate more paired data. We evaluate BIFI on two code repair datasets: GitHub-Python, a new dataset we introduce where the goal is to repair Python code with AST parse errors; and DeepFix, where the goal is to repair C code with compiler errors. BIFI outperforms existing methods, obtaining 90.5% repair accuracy on GitHub-Python (+28.5%) and 71.7% on DeepFix (+5.6%). Notably, BIFI does not require any labeled data; we hope it will be a strong starting point for unsupervised learning of various repair tasks.

cs.LG↗

arXiv2021

Swords: A Benchmark for Lexical Substitution with Improved Data Coverage and Quality

We release a new benchmark for lexical substitution, the task of finding appropriate substitutes for a target word in a context. To assist humans with writing, lexical substitution systems can suggest words that humans cannot easily think of. However, existing benchmarks depend on human recall as the only source of data, and therefore lack coverage of the substitutes that would be most helpful to humans. Furthermore, annotators often provide substitutes of low quality, which are not actually appropriate in the given context. We collect higher-coverage and higher-quality data by framing lexical substitution as a classification problem, guided by the intuition that it is easier for humans to judge the appropriateness of candidate substitutes than conjure them from memory. To this end, we use a context-free thesaurus to produce candidates and rely on human judgement to determine contextual appropriateness. Compared to the previous largest benchmark, our Swords benchmark has 4.1x more substitutes per target word for the same level of quality, and its substitutes are 1.5x more appropriate (based on human judgement) for the same number of substitutes.

cs.CL↗

arXiv2021

Selective Classification Can Magnify Disparities Across Groups

Selective classification, in which models can abstain on uncertain predictions, is a natural approach to improving accuracy in settings where errors are costly but abstentions are manageable. In this paper, we find that while selective classification can improve average accuracies, it can simultaneously magnify existing accuracy disparities between various groups within a population, especially in the presence of spurious correlations. We observe this behavior consistently across five vision and NLP datasets. Surprisingly, increasing abstentions can even decrease accuracies on some groups. To better understand this phenomenon, we study the margin distribution, which captures the model's confidences over all predictions. For symmetric margin distributions, we prove that whether selective classification monotonically improves or worsens accuracy is fully determined by the accuracy at full coverage (i.e., without any abstentions) and whether the distribution satisfies a property we call left-log-concavity. Our analysis also shows that selective classification tends to magnify full-coverage accuracy disparities. Motivated by our analysis, we train distributionally-robust models that achieve similar full-coverage accuracies across groups and show that selective classification uniformly improves each group on these models. Altogether, our results suggest that selective classification should be used with care and underscore the importance of training models to perform equally well across groups at full coverage.

cs.LG↗

arXiv2021

In-N-Out: Pre-Training and Self-Training using Auxiliary Information for Out-of-Distribution Robustness

Consider a prediction setting with few in-distribution labeled examples and many unlabeled examples both in- and out-of-distribution (OOD). The goal is to learn a model which performs well both in-distribution and OOD. In these settings, auxiliary information is often cheaply available for every input. How should we best leverage this auxiliary information for the prediction task? Empirically across three image and time-series datasets, and theoretically in a multi-task linear regression setting, we show that (i) using auxiliary information as input features improves in-distribution error but can hurt OOD error; but (ii) using auxiliary information as outputs of auxiliary pre-training tasks improves OOD error. To get the best of both worlds, we introduce In-N-Out, which first trains a model with auxiliary inputs and uses it to pseudolabel all the in-distribution inputs, then pre-trains a model on OOD auxiliary outputs and fine-tunes this model with the pseudolabels (self-training). We show both theoretically and empirically that In-N-Out outperforms auxiliary inputs or outputs alone on both in-distribution and OOD error.

cs.LG↗

arXiv2021

Beyond I.I.D.: Three Levels of Generalization for Question Answering on Knowledge Bases

Existing studies on question answering on knowledge bases (KBQA) mainly operate with the standard i.i.d assumption, i.e., training distribution over questions is the same as the test distribution. However, i.i.d may be neither reasonably achievable nor desirable on large-scale KBs because 1) true user distribution is hard to capture and 2) randomly sample training examples from the enormous space would be highly data-inefficient. Instead, we suggest that KBQA models should have three levels of built-in generalization: i.i.d, compositional, and zero-shot. To facilitate the development of KBQA models with stronger generalization, we construct and release a new large-scale, high-quality dataset with 64,331 questions, GrailQA, and provide evaluation settings for all three levels of generalization. In addition, we propose a novel BERT-based KBQA model. The combination of our dataset and model enables us to thoroughly examine and demonstrate, for the first time, the key role of pre-trained contextual embeddings like BERT in the generalization of KBQA.

cs.CL↗

arXiv2021

Task-Oriented Dialogue as Dataflow Synthesis

We describe an approach to task-oriented dialogue in which dialogue state is represented as a dataflow graph. A dialogue agent maps each user utterance to a program that extends this graph. Programs include metacomputation operators for reference and revision that reuse dataflow fragments from previous turns. Our graph-based state enables the expression and manipulation of complex user intents, and explicit metacomputation makes these intents easier for learned models to predict. We introduce a new dataset, SMCalFlow, featuring complex dialogues about events, weather, places, and people. Experiments show that dataflow graphs and metacomputation substantially improve representability and predictability in these natural dialogues. Additional experiments on the MultiWOZ dataset show that our dataflow representation enables an otherwise off-the-shelf sequence-to-sequence model to match the best existing task-specific state tracking model. The SMCalFlow dataset and code for replicating experiments are available at https://www.microsoft.com/en-us/research/project/dataflow-based-dialogue-semantic-machines.

cs.CL↗

arXiv2021

Prefix-Tuning: Optimizing Continuous Prompts for Generation

Fine-tuning is the de facto way to leverage large pretrained language models to perform downstream tasks. However, it modifies all the language model parameters and therefore necessitates storing a full copy for each task. In this paper, we propose prefix-tuning, a lightweight alternative to fine-tuning for natural language generation tasks, which keeps language model parameters frozen, but optimizes a small continuous task-specific vector (called the prefix). Prefix-tuning draws inspiration from prompting, allowing subsequent tokens to attend to this prefix as if it were "virtual tokens". We apply prefix-tuning to GPT-2 for table-to-text generation and to BART for summarization. We find that by learning only 0.1\% of the parameters, prefix-tuning obtains comparable performance in the full data setting, outperforms fine-tuning in low-data settings, and extrapolates better to examples with topics unseen during training.

cs.CL↗

arXiv2020

Understanding Black-box Predictions via Influence Functions

How can we explain the predictions of a black-box model? In this paper, we use influence functions -- a classic technique from robust statistics -- to trace a model's prediction through the learning algorithm and back to its training data, thereby identifying training points most responsible for a given prediction. To scale up influence functions to modern machine learning settings, we develop a simple, efficient implementation that requires only oracle access to gradients and Hessian-vector products. We show that even on non-convex and non-differentiable models where the theory breaks down, approximations to influence functions can still provide valuable information. On linear models and convolutional neural networks, we demonstrate that influence functions are useful for multiple purposes: understanding model behavior, debugging models, detecting dataset errors, and even creating visually-indistinguishable training-set attacks.

stat.ML↗

arXiv2020

Concept Bottleneck Models

We seek to learn models that we can interact with using high-level concepts: if the model did not think there was a bone spur in the x-ray, would it still predict severe arthritis? State-of-the-art models today do not typically support the manipulation of concepts like "the existence of bone spurs", as they are trained end-to-end to go directly from raw input (e.g., pixels) to output (e.g., arthritis severity). We revisit the classic idea of first predicting concepts that are provided at training time, and then using these concepts to predict the label. By construction, we can intervene on these concept bottleneck models by editing their predicted concept values and propagating these changes to the final prediction. On x-ray grading and bird identification, concept bottleneck models achieve competitive accuracy with standard end-to-end models, while enabling interpretation in terms of high-level clinical concepts ("bone spurs") or bird attributes ("wing color"). These models also allow for richer human-model interaction: accuracy improves significantly if we can correct model mistakes on concepts at test time.

cs.LG↗

arXiv2020

Removing Spurious Features can Hurt Accuracy and Affect Groups Disproportionately

The presence of spurious features interferes with the goal of obtaining robust models that perform well across many groups within the population. A natural remedy is to remove spurious features from the model. However, in this work we show that removal of spurious features can decrease accuracy due to the inductive biases of overparameterized models. We completely characterize how the removal of spurious features affects accuracy across different groups (more generally, test distributions) in noiseless overparameterized linear regression. In addition, we show that removal of spurious feature can decrease the accuracy even in balanced datasets -- each target co-occurs equally with each spurious feature; and it can inadvertently make the model more susceptible to other spurious features. Finally, we show that robust self-training can remove spurious features without affecting the overall accuracy. Experiments on the Toxic-Comment-Detectoin and CelebA datasets show that our results hold in non-linear models.

cs.LG↗

arXiv2020

Feature Noise Induces Loss Discrepancy Across Groups

The performance of standard learning procedures has been observed to differ widely across groups. Recent studies usually attribute this loss discrepancy to an information deficiency for one group (e.g., one group has less data). In this work, we point to a more subtle source of loss discrepancy---feature noise. Our main result is that even when there is no information deficiency specific to one group (e.g., both groups have infinite data), adding the same amount of feature noise to all individuals leads to loss discrepancy. For linear regression, we thoroughly characterize the effect of feature noise on loss discrepancy in terms of the amount of noise, the difference between moments of the two groups, and whether group information is used or not. We then show this loss discrepancy does not vanish immediately if a shift in distribution causes the groups to have similar moments. On three real-world datasets, we show feature noise increases the loss discrepancy if groups have different distributions, while it does not affect the loss discrepancy on datasets where groups have similar distributions.

cs.LG↗

arXiv2020

Enabling certification of verification-agnostic networks via memory-efficient semidefinite programming

Convex relaxations have emerged as a promising approach for verifying desirable properties of neural networks like robustness to adversarial perturbations. Widely used Linear Programming (LP) relaxations only work well when networks are trained to facilitate verification. This precludes applications that involve verification-agnostic networks, i.e., networks not specially trained for verification. On the other hand, semidefinite programming (SDP) relaxations have successfully be applied to verification-agnostic networks, but do not currently scale beyond small networks due to poor time and space asymptotics. 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. By exploiting iterative eigenvector methods, we express all solver operations in terms of forward and backward passes through the network, enabling efficient use of hardware like GPUs/TPUs. For two verification-agnostic networks on MNIST and CIFAR-10, we significantly improve L-inf verified robust accuracy from 1% to 88% and 6% to 40% respectively. We also demonstrate tight verification of a quadratic stability specification for the decoder of a variational autoencoder.

cs.LG↗

arXiv2020

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. Defenses based on regularization and adversarial training have been proposed, but often followed by new, stronger attacks that defeat these defenses. Can we somehow end this arms race? In this work, we study this problem for neural networks with one hidden layer. We first propose a method based on a semidefinite relaxation that outputs a certificate that for a given network and test input, no attack can force the error to exceed a certain value. Second, as this certificate is differentiable, we jointly optimize it with the network parameters, providing an adaptive regularizer that encourages robustness against all attacks. On MNIST, our approach produces a network and a certificate that no attack that perturbs each pixel by at most ε= 0.1 can cause more than 35% test error.

cs.LG↗