Related papers: Learning from Negative Examples: Why Warning-Frame…

It has recently been demonstrated empirically that in-context learning emerges in transformers when certain distributional properties are present in the training data, but this ability can also diminish upon further training. We provide a…

Learning with limited data is one of the biggest problems of machine learning. Current approaches to this issue consist in learning general representations from huge amounts of data before fine-tuning the model on a small dataset of…

A self-adaptive system can modify its own structure and behavior at runtime based on its perception of the environment, of itself and of its requirements. To develop a self-adaptive system, software developers codify knowledge about the…

Regression models usually tend to recover a noisy signal in the form of a combination of regressors, also called features in machine learning, themselves being the result of a learning process.The alignment of the prior covariance feature…

Prevailing alignment methods induce opaque parameter changes, obscuring what models truly learn. To address this, we introduce Feature Steering with Reinforcement Learning (FSRL), a framework that trains a lightweight adapter to steer model…

Federated learning allows multiple users to collaboratively train a shared classification model while preserving data privacy. This approach, where model updates are aggregated by a central server, was shown to be vulnerable to poisoning…

Large language models have demonstrated strong capabilities to learn in-context, where exemplar input-output pairings are appended to the prompt for demonstration. However, existing work has demonstrated the ability of models to learn…

Why do language models trained on contradictory data prefer correct answers? In controlled experiments with small transformers (3.5M--86M parameters), we show that this preference tracks the compressibility structure of errors rather than…

Despite the increasing prevalence of large language models (LLMs), we still have a limited understanding of how their representational spaces are structured. This limits our ability to interpret how and what they learn or relate them to…

In naturalistic learning problems, a model's input contains a wide range of features, some useful for the task at hand, and others not. Of the useful features, which ones does the model use? Of the task-irrelevant features, which ones does…

A lot of recent advances in unsupervised feature learning are based on designing features which are invariant under semantic data augmentations. A common way to do this is contrastive learning, which uses positive and negative samples. Some…

Reinforcement Learning (RL) environments can produce training data with spurious correlations between features due to the amount of training data or its limited feature coverage. This can lead to RL agents encoding these misleading…

We investigate the extent to which modern, neural language models are susceptible to structural priming, the phenomenon whereby the structure of a sentence makes the same structure more probable in a follow-up sentence. We explore how…

Text representations learned by machine learning models often encode undesirable demographic information of the user. Predictive models based on these representations can rely on such information, resulting in biased decisions. We present a…

Recent studies on backdoor attacks in model training have shown that polluting a small portion of training data is sufficient to produce incorrect manipulated predictions on poisoned test-time data while maintaining high clean accuracy in…

Existing works have shown that fine-tuned textual transformer models achieve state-of-the-art prediction performances but are also vulnerable to adversarial text perturbations. Traditional adversarial evaluation is often done \textit{only…

Shortcut learning, or `Clever Hans effect` refers to situations where a learning agent (e.g., deep neural networks) learns spurious correlations present in data, resulting in biased models. We focus on finding shortcuts in deep learning…

We introduce a framework for unsupervised learning of structured predictors with overlapping, global features. Each input's latent representation is predicted conditional on the observable data using a feature-rich conditional random field.…

Empirical studies suggest that machine learning models often rely on features, such as the background, that may be spuriously correlated with the label only during training time, resulting in poor accuracy during test-time. In this work, we…

Standard language model training employs gold human documents or human-human interaction data, and treats all training data as positive examples. Growing evidence shows that even with very large amounts of positive training data, issues…