As language models (LMs) are widely utilized in personalized communication scenarios (e.g., sending emails, writing social media posts) and endowed with a certain level of agency, ensuring they act in accordance with the contextual privacy norms becomes increasingly critical. However, quantifying the privacy norm awareness of LMs and the emerging privacy risk in LM-mediated communication is challenging due to (1) the contextual and long-tailed nature of privacy-sensitive cases, and (2) the lack of evaluation approaches that capture realistic application scenarios. To address these challenges, we propose PrivacyLens, a novel framework designed to extend privacy-sensitive seeds into expressive vignettes and further into agent trajectories, enabling multi-level evaluation of privacy leakage in LM agents' actions. We instantiate PrivacyLens with a collection of privacy norms grounded in privacy literature and crowdsourced seeds. Using this dataset, we reveal a discrepancy between LM performance in answering probing questions and their actual behavior when executing user instructions in an agent setup. State-of-the-art LMs, like GPT-4 and Llama-3-70B, leak sensitive information in 25.68% and 38.69% of cases, even when prompted with privacy-enhancing instructions. We also demonstrate the dynamic nature of PrivacyLens by extending each seed into multiple trajectories to red-team LM privacy leakage risk.

Deep learning models are now trained on increasingly larger datasets, making it crucial to reduce computational costs and improve data quality. Dataset distillation aims to distill a large dataset into a small synthesized dataset such that models trained on it can achieve similar performance to those trained on the original dataset. While there have been many empirical efforts to improve dataset distillation algorithms, a thorough theoretical analysis and provable, efficient algorithms are still lacking. In this paper, by focusing on dataset distillation for kernel ridge regression (KRR), we show that one data point per class is already necessary and sufficient to recover the original model's performance in many settings. For linear ridge regression and KRR with surjective feature mappings, we provide necessary and sufficient conditions for the distilled dataset to recover the original model's parameters. For KRR with injective feature mappings of deep neural networks, we show that while one data point per class is not sufficient in general, $k+1$ data points can be sufficient for deep linear neural networks, where $k$ is the number of classes. Our theoretical results enable directly constructing analytical solutions for distilled datasets, resulting in a provable and efficient dataset distillation algorithm for KRR. We verify our theory experimentally and show that our algorithm outperforms previous work such as KIP while being significantly more efficient, e.g.

In recent years, transformer-based models have revolutionized deep learning, particularly in sequence modeling. To better understand this phenomenon, there is a growing interest in using Markov input processes to study transformers. However, our current understanding in this regard remains limited with many fundamental questions about how transformers learn Markov chains still unanswered. In this paper, we address this by focusing on first-order Markov chains and single-layer transformers, providing a comprehensive characterization of the learning dynamics in this context. Specifically, we prove that transformer parameters trained on next-token prediction loss can either converge to global or local minima, contingent on the initialization and the Markovian data properties, and we characterize the precise conditions under which this occurs. To the best of our knowledge, this is the first result of its kind highlighting the role of initialization. We further demonstrate that our theoretical findings are corroborated by empirical evidence. Based on these insights, we provide guidelines for the initialization of single-layer transformers and demonstrate their effectiveness. Finally, we outline several open problems in this arena.

Evidential Deep Learning (EDL), grounded in Evidence Theory and Subjective Logic (SL), provides a robust framework to estimate uncertainty for out-of-distribution (OOD) detection alongside traditional classification probabilities.However, the EDL framework is constrained by its focus on evidence that supports only single categories, neglecting the other collective evidences that could corroborate multiple in-distribution categories. This limitation leads to a diminished estimation of uncertainty and a subsequent decline in OOD detection performance.Additionally, EDL encounters the vanishing gradient problem within its fully-connected layers, further degrading classification accuracy.To address these issues, we introduce hyper-domain and propose Hyper-opinion Evidential Deep Learning (HEDL). HEDL extends the evidence modeling paradigm by explicitly integrating sharp evidence, which supports a singular category, with vague evidence that accommodates multiple potential categories.Additionally, we propose a novel opinion projection mechanism that translates hyper-opinion into multinomial-opinion, which is then optimized within the EDL framework to ensure precise classification and refined uncertainty estimation.HEDL integrates evidences across various categories to yield a holistic evidentiary foundation for achieving superior OOD detection. Furthermore, our proposed opinion projection method effectively mitigates the vanishing gradient issue, ensuring classification accuracy without additional model complexity. Extensive experiments over many datasets demonstrate our proposed method outperforms existing OOD detection methods.

Co-crystallization is an accessible way to control physicochemical characteristics of organic crystals, which finds many biomedical applications. In this work, we present Generative Method for Co-crystal Design (GEMCODE), a novel pipeline for automated co-crystal screening based on the hybridization of deep generative models and evolutionary optimization for broader exploration of the target chemical space. GEMCODE enables fast co-crystal design with target tabletability profiles, which is crucial for the development of pharmaceuticals. With a series of experimental studies highlighting validation and discovery cases, we show that GEMCODE is effective even under realistic computational constraints. Furthermore, we explore the potential of language models in generating co-crystals. Finally, we present numerous previously unknown co-crystals predicted by GEMCODE and discuss its potential in accelerating drug development.

The growing safety concerns surrounding large language models raise an urgent need to align them with diverse human preferences to simultaneously enhance their helpfulness and safety. A promising approach is to enforce safety constraints through Reinforcement Learning from Human Feedback (RLHF). For such constrained RLHF, typical Lagrangian-based primal-dual policy optimization methods are computationally expensive and often unstable. This paper presents a perspective of dualization that reduces constrained alignment to an equivalent unconstrained alignment problem. We do so by pre-optimizing a smooth and convex dual function that has a closed form. This shortcut eliminates the need for cumbersome primal-dual policy iterations, greatly reducing the computational burden and improving training stability. Our strategy leads to two practical algorithms in model-based and preference-based settings (MoCAN and PeCAN, respectively). A broad range of experiments demonstrate the effectiveness and merits of our algorithms.

While the expected calibration error (ECE), which employs binning, is widely adopted to evaluate the calibration performance of machine learning models, theoretical understanding of its estimation bias is limited. In this paper, we present the first comprehensive analysis of the estimation bias in the two common binning strategies, uniform mass and uniform width binning.Our analysis establishes upper bounds on the bias, achieving an improved convergence rate. Moreover, our bounds reveal, for the first time, the optimal number of bins to minimize the estimation bias. We further extend our bias analysis to generalization error analysis based on the information-theoretic approach, deriving upper bounds that enable the numerical evaluation of how small the ECE is for unknown data. Experiments using deep learning models show that our bounds are nonvacuous thanks to this information-theoretic generalization analysis approach.

Scaling law that rewards large datasets, complex models and enhanced data granularity has been observed in various fields of deep learning. Yet, studies on time series forecasting have cast doubt on scaling behaviors of deep learning methods for time series forecasting: while more training data improves performance, more capable models do not always outperform less capable models, and longer input horizon may hurt performance for some models. We propose a theory for scaling law for time series forecasting that can explain these seemingly abnormal behaviors. We take into account the impact of dataset size and model complexity, as well as time series data granularity, particularly focusing on the look-back horizon, an aspect that has been unexplored in previous theories. Furthermore, we empirically evaluate various models using a diverse set of time series forecasting datasets, which (1) verifies the validity of scaling law on dataset size and model complexity within the realm of time series forecasting, and (2) validates our theoretical framework, particularly regarding the influence of look back horizon. We hope our findings may inspire new models targeting time series forecasting datasets of limited size, as well as large foundational datasets and models for time series forecasting in future works.

We investigate Reinforcement Learning from Human Feedback (RLHF) in the context of a general preference oracle. In particular, we do not assume the existence of a reward function and an oracle preference signal drawn from the Bradley-Terry model as most of the prior works do. We consider a standard mathematical formulation, the reverse-KL regularized minimax game between two LLMs for RLHF under general preference oracle. The learning objective of this formulation is to find a policy so that it is consistently preferred by the KL-regularized preference oracle over any competing LLMs. We show that this framework is strictly more general than the reward-based one, and propose sample-efficient algorithms for both the offline learning from a pre-collected preference dataset and online learning where we can query the preference oracle along the way of training. Empirical studies verify the effectiveness of the proposed framework.

Reinforcement Learning from Human Feedback (RLHF) has been pivotal in aligning Large Language Models with human values but often suffers from overoptimization due to its reliance on a proxy reward model. To mitigate this limitation, we first propose a lightweight uncertainty quantification method that assesses the reliability of the proxy reward using only the last layer embeddings of the reward model. Enabled by this efficient uncertainty quantification method, we formulate AdvPO, a distributionally robust optimization procedure to tackle the reward overoptimization problem in RLHF. Through extensive experiments on the Anthropic HH and TL;DR summarization datasets, we verify the effectiveness of AdvPO in mitigating the overoptimization problem, resulting in enhanced RLHF performance as evaluated through human-assisted evaluation.