论文检索

Theory · Optimization

Jinyang ZHENG, Tong Wu

Numerous studies on low-rank adaptation (LoRA) emerged in recent years, with the aim of accelerating the convergence of the LoRA framework. In this paper, we leverage the horizontal lift theory from differential geometry to establish the general iteration scheme on the quotient manifold \mathbb{R}\_\*^{m \times r} \times \mathbb{R}\_\*^{n \times r}/\sim. By endowing the LoRA framework with Riemannian quotient geometries, our theory not only guarantees efficient feature learning but also bridges the LoRA algorithms and the pre-training algorithms for large models. Furthermore, we theoretically analyze the role of the weight decay matrix $\epsilon_{decay}I$ in efficient feature learning and then replace it with the Sylvester matrix $K$, indicating that the theory helps remove an important hyperparameter while generating accurate and computationally efficient optimizers. Based on the general scheme, we propose two efficient LoRA optimizers with runtime analysis, Adam-Sylvester (AdamS) and LRACS, then conduct experiments on the transformer-based networks. The results demonstrate evident improvements over existing optimizers.

Deep Learning · Attention Mechanisms

Hariprasath Govindarajan, Per Sidén, Jacob Roll, Fredrik Lindsten

The Transformer model architecture has become one of the most widely used in deep learning and the attention mechanism is at its core. The standard attention formulation uses a softmax operation applied to a scaled dot product between query and key vectors. We explore the role played by norms of the queries and keys, which can cause training instabilities when they arbitrarily increase. We demonstrate how this can happen even in simple Transformer models, in the presence of easy-to-learn spurious patterns in the data. We propose a new attention formulation, QUEry-modulated Spherical aTtention (QUEST), that constrains the keys to a hyperspherical latent space, while still allowing individual tokens to flexibly control the sharpness of the attention distribution. QUEST can be easily used as a drop-in replacement for standard attention. We focus on vision applications while also exploring other domains to highlight the method's generality. We show that (1) QUEST trains without instabilities and (2) produces models with improved performance (3) that are robust to data corruptions and adversarial attacks.

Social Aspects · Privacy-preserving Statistics and Machine Learning

Aurélien Bellet, Edwige Cyffers, Davide Frey, Romaric Gaudel, Dimitri Lerévérend, Francois Taiani

Decentralized Learning (DL) enables users to collaboratively train models without sharing raw data by iteratively averaging local updates with neighbors in a network graph. This setting is increasingly popular for its scalability and its ability to keep data local under user control. Strong privacy guarantees in DL are typically achieved through Differential Privacy (DP), with results showing that DL can even amplify privacy by disseminating noise across peer-to-peer communications. Yet in practice, the observed privacy-utility trade-off often appears worse than in centralized training, which may be due to limitations in current DP accounting methods for DL. In this paper, we show that recent advances in centralized DP accounting based on Matrix Factorization (MF) for analyzing temporal noise correlations can also be leveraged in DL. By generalizing existing MF results, we show how to cast both standard DL algorithms and common trust models into a unified formulation. This yields tighter privacy accounting for existing DP-DL algorithms and provides a principled way to develop new ones. To demonstrate the approach, we introduce MAFALDA-SGD, a gossip-based DL algorithm with user-level correlated noise that outperforms existing methods on synthetic and real-world graphs.

Reinforcement Learning · Everything Else

Anthony GX-Chen, Jatin Prakash, Jeff Guo, Rob Fergus, Rajesh Ranganath

Classical intuitions cast minimizing reverse KL as "mode seeking" and forward KL as "mass covering". In KL-regularized reinforcement learning, however, the regularizer determines _both_ the target distribution's shape _and_ the divergence being implicitly minimized, making its role more nuanced than simply inducing generic mode-seeking or mass-covering behaviour. Specifically, the target distribution is defined jointly by the reward function, the reference model, the type of regularizer, and the regularization strength. We show that under common settings—such as low regularization strength and equal verifiable rewards—both forward and reverse KL regularization tend to specify target distributions whose mass concentrates on a single high-reward region. Thus, the objective itself _by construction_ induces diversity collapse, regardless of the policy optimization algorithm used. Building on this perspective, we introduce a simple and scalable modification that rescales rewards to induce target distributions assigning substantial probability across _all_ high-reward regions. This yields a principled objective that maintains high solution quality while achieving broad reward-mode coverage. Empirically, this approach improves post-training diversity and performance for Large Language Models and Chemical Language Models, and is effective with either forward or reverse KL regularization, while using either naively fails.

Optimization · Sampling and Optimization

Advait Parulekar, Litu Rout, Karthikeyan Shanmugam, Sanjay Shakkottai

We study the problem of posterior sampling in the context of score based generative models. We have a trained score network for a prior $p(x)$, a measurement model $p(y|x)$, and are tasked with sampling from the posterior $p(x|y)$. Prior work has shown this to be intractable in KL (in the worst case) under well-accepted computational hardness assumptions. Despite this, popular algorithms for tasks such as image super-resolution, stylization, and reconstruction enjoy empirical success. Rather than establishing distributional assumptions or restricted settings under which exact posterior sampling is tractable, we view this as a more general "tilting" problem of biasing a distribution towards a measurement. Under minimal assumptions, we show that one can tractably sample from a distribution that is simultaneously close to the posterior of a noised prior in KL divergence and the true posterior in Fisher divergence. Intuitively, this combination ensures that the resulting sample is consistent with both the measurement and the prior. To the best of our knowledge these are the first formal results for (approximate) posterior sampling in polynomial time.

Theory · Learning Theory

Anna Scampicchio, Leonardo Felipe Toso, Rahel Rickenbach, James Anderson, Melanie Zeilinger

A major challenge in physics-informed machine learning is to understand how the incorporation of prior domain knowledge affects learning rates when data are dependent. Focusing on empirical risk minimization with physics-informed regularization, we derive complexity-dependent bounds on the excess risk in probability and in expectation. We prove that, when the physical prior information is aligned, the learning rate improves from the (slow) Sobolev minimax rate to the (fast) optimal i.i.d. one without sample-size deflation due to data dependence.

Optimization · Learning for Optimization

Tianyou Li, Haijun Zou, JIAYUAN WU, Zaiwen Wen

Routing problems are canonical combinatorial optimization tasks with wide-ranging applications in logistics, transportation, and supply chain management. However, solving these problems becomes significantly more challenging when complex constraints are involved. In this paper, we propose LMask, a novel learning framework that utilizes dynamic masking to generate high-quality feasible solutions for constrained routing problems. LMask introduces the LazyMask decoding method, which lazily refines feasibility masks with the backtracking mechanism. In addition, it employs the refinement intensity embedding to encode the search trace into the model, mitigating representation ambiguities induced by backtracking. To further reduce sampling cost, LMask sets a backtracking budget during decoding, while constraint violations are penalized in the loss function during training to counteract infeasibility caused by this budget. We provide theoretical guarantees for the validity and probabilistic optimality of our approach. Extensive experiments on the traveling salesman problem with time windows (TSPTW) and TSP with draft limits (TSPDL) demonstrate that LMask achieves state-of-the-art feasibility rates and solution quality, outperforming existing neural methods.

Applications · Genetics, Cell Biology, Health, etc

Yize Zhou, Haorui Li, Shengchao Liu

Chromosomes are the carriers of genetic information. Further understanding their 3D structure can help reveal gene-regulatory mechanisms and cellular functions. However, high-resolution 3D structures are often missing due to the high cost and inherent noise of experimental screening. A standard pipeline for reconstructing the chromosome 3D structure first applies the single-cell Hi-C high-throughput screening method to measure pairwise interactions between DNA fragments at different resolutions; then it adopts computational methods to reconstruct the 3D structures from these contacts. These include traditional numerical methods and deep learning models, which struggle with limited model expressiveness and poor generalization across resolutions. To handle this issue, we propose InertialGenome, a novel transformer-based framework for robust and resolution-agnostic chromosome reconstruction. InertialGenome first adopts the inertial frame for the pose canonicalization. Then, based on such an invariant pose, it proposes a Transformer with geometry-aware positional encoding, leveraging Nyström estimation. To verify the effectiveness of InertialGenome, we conduct experiments on two single-cell 3D reconstruction datasets with four resolutions, reaching superior performance over all four computational baselines. Additionally, we observe that the 3D structure reconstructed by InertialGenome is more in line with the results of real experimental results on two functional verification tasks. Finally, we leverage InertialGenome for cross-resolution transfer learning, yielding up to a 5\% improvement from low to high resolution. The source code is available at https://github.com/yize1203/InertialGenome.

Computer Vision · 3D Rendering & Reconstruction

Min-Seop Kwak, Junho Kim, Sangdoo Yun, Dongyoon Han, Taekyung Kim, Seungryong Kim, Jin-Hwa Kim

We introduce a diffusion-based framework that generates aligned novel view images and geometries via a warping‐and‐inpainting methodology. Unlike prior methods that require dense posed images or pose-embedded generative models limited to in‐domain views, our method leverages off‐the‐shelf geometry predictors to predict partial geometries viewed from reference images, and formulates novel view synthesis as an inpainting task for both image and geometry. To ensure accurate alignment between the generated image and geometry, we propose cross-modal attention instillation where the attention maps from an image diffusion branch are injected into a parallel geometry diffusion branch during both training and inference. This multi-task approach achieves synergistic effects, facilitating both geometrically robust image synthesis and geometry prediction. We further introduce proximity‐based mesh conditioning to integrate depth and normal cues, interpolating between point cloud and filtering erroneously predicted geometry from influencing the generation process. Empirically, our method achieves high-fidelity extrapolative view synthesis, delivers competitive reconstruction under interpolation settings, and produces geometrically aligned point clouds as 3D completion.

Applications · Neuroscience, Cognitive Science

Bahti Zakirov, Gasper Tkacik

Normative and task-driven theories offer powerful top-down explanations for biological systems, yet the goals of quantitatively arbitrating between competing theories, and utilizing them as inductive biases to improve data-driven fits of real biological datasets are prohibitively laborious, and often impossible. To this end, we introduce a Bayesian meta-learning framework designed to automatically convert raw functional predictions from normative theories into tractable probabilistic models. We employ adaptive deep kernel Gaussian processes, meta-learning a kernel on synthetic data generated from a normative theory. This Theory-Informed Kernel specifies a probabilistic model representing the predictions of a given theoretical model: usable for both fitting data and rigorously validating the theory. As a demonstration, we apply our framework to the early visual system, using efficient coding as our normative theory. We show improved response prediction accuracy in ex vivo recordings of mouse retinal ganglion cells stimulated by natural scenes compared to conventional data-driven baselines, while providing accurate uncertainty estimates and interpretable representations. Using exact Bayesian model selection, we also show that our informed kernel can accurately infer the degree of theory-match from data, confirming faithful encapsulation of theory structure. This work provides a more general, scalable, and automated approach for integrating theoretical knowledge into data-driven scientific inquiry in neuroscience and beyond.

Reinforcement Learning · Deep RL

Chongyi Zheng, Mahsa Bastankhah, Grace Liu, Benjamin Eysenbach

Goal-conditioned reinforcement learning learns to reach goals instead of optimizing hand-crafted rewards. Despite its popularity, the community often categorizes goal-conditioned reinforcement learning as a special case of reinforcement learning. In this post, we aim to build a direct conversion from any reward-maximization reinforcement learning problem to a goal-conditioned reinforcement learning problem, and to draw connections with the stochastic shortest path framework. Our conversion provides a new perspective on the reinforcement learning problem: *maximizing rewards is equivalent to reaching some goals*.

Theory · Interpretability and Visualization

Yavuz Faruk Bakman, Sungmin Kang, Zhiqi Huang, Duygu Nur Yaldiz, Catarina Belém, Chenyang Zhu, Anoop Kumar, Alfy Samuel, Daben Liu, Salman Avestimehr 等

Uncertainty Quantification (UQ) research has primarily focused on closed-book factual question answering (QA), while contextual QA remains unexplored, despite its importance in real-world applications. In this work, we focus on UQ for the contextual QA task and propose a theoretically grounded approach to quantify \emph{epistemic uncertainty}. We begin by introducing a task-agnostic, token-level uncertainty measure defined as the cross-entropy between the predictive distribution of the given model and the unknown true distribution. By decomposing this measure, we isolate the epistemic component and approximate the true distribution by a perfectly prompted, idealized model. We then derive an upper bound for epistemic uncertainty and show that it can be interpreted as semantic feature gaps in the given model’s hidden representations relative to the ideal model. We further apply this generic framework to the contextual QA task and hypothesize that three features approximate this gap: \emph{context-reliance} (using the provided context rather than parametric knowledge), \emph{context comprehension} (extracting relevant information from context), and \emph{honesty} (avoiding intentional lies). Using a top-down interpretability approach, we extract these features by using only a small number of labeled samples and ensemble them to form a robust uncertainty score. Experiments on multiple QA benchmarks in both in-distribution and out-of-distribution settings show that our method substantially outperforms state-of-the-art unsupervised (sampling-free and sampling-based) and supervised UQ methods, achieving up to a 13-point PRR improvement while incurring a negligible inference overhead. The code is available at https://github.com/Ybakman/Feature-Gaps.

Applications · Genetics, Cell Biology, Health, etc

Rohit Dilip, Evan Zhang, Ayush Varshney, David Van Valen

Protein structure tokenizers enable the creation of multimodal models of protein structure, sequence, and function. Current approaches to protein structure tokenization rely on bespoke components that are invariant to spatial symmetries, but that are challenging to optimize and scale. We present Kanzi, a flow-based tokenizer for tokenization and generation of protein structures. Kanzi consists of a diffusion autoencoder trained with a flow matching loss. We show that this approach simplifies several aspects of protein structure tokenizers: frame-based representations can be replaced with global coordinates, complex losses are replaced with a single flow matching loss, and SE(3)-invariant attention operations can be replaced with standard attention. We find that these changes stabilize the training of parameter-efficient models that outperform existing tokenizers on reconstruction metrics at a fraction of the model size and training cost. An autoregressive model trained with Kanzi outperforms similar generative models that operate over tokens, although it does not yet match the performance of state-of-the-art continuous diffusion models.

Reinforcement Learning · Deep RL

Chongyi Zheng, Seohong Park, Sergey Levine, Benjamin Eysenbach

Large-scale pre-training has fundamentally changed how machine learning research is done today: large foundation models are trained once, and then can be used by anyone in the community (including those without data or compute resources to train a model from scratch) to adapt and fine-tune to specific tasks. Applying this same framework to reinforcement learning (RL) is appealing because it offers compelling avenues for addressing core challenges in RL, including sample efficiency and robustness. However, there remains a fundamental challenge to pre-train large models in the context of RL: actions have long-term dependencies, so training a foundation model that reasons across *time* is important. Recent advances in generative AI have provided new tools for modeling highly complex distributions. In this paper, we build a probabilistic model to predict which states an agent will visit in the temporally distant future (i.e., an occupancy measure) using flow matching. As large datasets are often constructed by many distinct users performing distinct tasks, we include in our model a latent variable capturing the user intention. This intention increases the expressivity of our model, and enables adaptation with generalized policy improvement. We call our proposed method **intention-conditioned flow occupancy models (InFOM)**. Comparing with alternative methods for pre-training, our experiments on $36$ state-based and $4$ image-based benchmark tasks demonstrate that the proposed method achieves $1.8 \times$ median improvement in returns and increases success rates by $36\\%$.

Jiawen Zhang, Lipeng He, Kejia Chen, Jian Lou, Jian Liu, Xiaohu Yang, Ruoxi Jia

Fine-tuning safety-aligned large language models (LLMs) can substantially compromise their safety. Previous approaches require many safety samples or calibration sets, which not only incur significant computational overhead during realignment but also lead to noticeable degradation in model utility. Contrary to this belief, we show that safety alignment can be fully recovered with only a single safety example, without sacrificing utility and at minimal cost. Remarkably, this recovery is effective regardless of the number of harmful examples used in fine-tuning or the size of the underlying model, and convergence is achieved within just a few epochs. Furthermore, we uncover the low-rank structure of the safety gradient, which explains why such efficient correction is possible. We validate our findings across five safety-aligned LLMs and multiple datasets, demonstrating the generality of our approach.

Social Aspects · Trustworthy Machine Learning

Thibaud Gloaguen, Mark Vero, Robin Staab, Martin Vechev

Finetuning open-weight Large Language Models (LLMs) is standard practice for achieving task-specific performance improvements. Until now, finetuning has been regarded as a controlled and secure process in which training on benign datasets leads to predictable behaviors. In this paper, we demonstrate, for the first time, that an adversary can create compromised LLMs that are performant and benign, yet exhibit adversarial behaviors once finetuned by downstream users. To this end, we propose an attack, FAB (Finetuning-activated Adversarial Behaviors), which compromises an LLM via meta-learning techniques that simulate downstream finetuning, explicitly optimizing for the emergence of adversarial behaviors in the finetuned models. At the same time, the compromised LLM is regularized to retain general capabilities and to exhibit no adversarial behaviors prior to finetuning. As a result, when users finetune (e.g., instruction-tuning, distillation, DPO) the seemingly benign model on their own datasets, they unknowingly trigger its dormant adversarial behavior. We experimentally demonstrate the effectiveness of FAB across multiple LLMs and three commonly considered target behaviors: unsolicited advertising, jailbreakability, and over-refusal. We show that FAB-triggers are robust to various finetuning choices made by the user (e.g., dataset, number of steps, scheduler, post-training algorithm). Our findings challenge prevailing assumptions on the security of finetuning, revealing a critical attack vector.

Social Aspects · Trustworthy Machine Learning

Mengxuan Hu, Vivek Datla, Anoop Kumar, Zihan Guan, Sheng Li, Alfy Samuel, Daben Liu

Recent advances in alignment techniques such as Supervised Fine-Tuning (SFT), Reinforcement Learning from Human Feedback (RLHF), and Direct Preference Optimization (DPO) have improved the safety of large language models (LLMs). However, these LLMs remain vulnerable to jailbreak attacks that disguise harmful intent through indirect or deceptive phrasing. Using causal intervention, we empirically demonstrate that this vulnerability stems from shallow alignment mechanisms that lack deep reasoning, often rejecting harmful prompts without truly understanding why they are harmful. To mitigate this vulnerability, we propose enhancing alignment through reasoning-aware post-training. We construct and release a novel Chain-of-Thought (CoT) fine-tuning dataset that includes both utility-oriented and safety-critical prompts with step-by-step rationales. Fine-tuning on this dataset encourages models to produce principled refusals grounded in reasoning, outperforming standard SFT baselines. Furthermore, inspired by failure patterns in CoT fine-tuning, we introduce **Alignment-Weighted DPO**, which targets the most problematic parts of an output by assigning different preference weights to the reasoning and final-answer segments. This produces finer-grained, targeted updates than vanilla DPO and improves robustness to diverse jailbreak strategies. Extensive experiments across multiple safety and utility benchmarks show that our method consistently improves alignment robustness while maintaining overall model utility.

Applications · Chemistry and Drug Discovery

Sai Pooja Mahajan, Natasa Tagasovska, Stefania Vasilaki, Arian Jamasb, Andrew Watkins, Rajesh Ranganath

Improving antibody binding to an antigen without antibody–antigen complex structures or antigen-specific training data is a central challenge in therapeutic protein design. We introduce **AffinityEnhancer**, a framework for one-shot antibody affinity improvement with strong generalization: given a single lead sequence, we propose variants that increase affinity without fine-tuning on the lead and without using antigen information, epitope/paratope labels, or the lead’s structure in complex with the antigen. During training, AffinityEnhancer leverages a pan-antigen dataset of diverse binding environments (antigens) and constructs paired examples of related sequences with higher vs. lower measured binding. A shared, structure-aware module learns to transform low-affinity sequences toward high-affinity ones, distilling consistent, causal features associated with improved binding across environments. By combining pretrained sequence–structure embeddings with a sequence decoder, AffinityEnhancer generalizes to entirely unseen antibody seeds. Across multiple held-out internal and public leads, AffinityEnhancer concentrates mutations on the rim of the paratope, outperforms existing structure-conditioned and inpainting baselines, and achieves substantial in silico affinity gains in true one-shot experiments, despite never observing antigen-specific data at test time.

General Machine Learning · Everything Else

Kejing Lu, Chuan Xiao, Yoshiharu Ishikawa

In this paper, we study the angle testing problem in the context of similarity search in high-dimensional Euclidean spaces and propose two projection-based probabilistic kernel functions, one designed for angle comparison and the other for angle thresholding. Unlike existing approaches that rely on random projection vectors drawn from Gaussian distributions, our approach leverages reference angles and adopts a deterministic structure for the projection vectors. Notably, our kernel functions do not require asymptotic assumptions, such as the number of projection vectors tending to infinity, and can be theoretically and experimentally shown to outperform Gaussian-distribution-based kernel functions. We apply the proposed kernel function to Approximate Nearest Neighbor Search (ANNS) and demonstrate that our approach achieves a 2.5x--3x higher query-per-second (QPS) throughput compared to the widely-used graph-based search algorithm HNSW. Our code and data are available at https://github.com/KejingLu-810/KS.

Deep Learning · Generative Models and Autoencoders

Linghao Kong, Angelina Ning, Micah Adler, Nir Shavit

Modern large language models increasingly use smooth activation functions such as GELU or SiLU, allowing negative pre-activations to carry both signal and gradient. Nevertheless, many neuron-level interpretability analyses have historically focused on large positive activations, often implicitly treating the negative region as less informative, a carryover from the ReLU-era. We challenge this assumption and ask whether and how negative pre-activations are leveraged by models. We address this question by studying a sparse subpopulation of Wasserstein neurons whose output distributions deviate strongly from a Gaussian baseline and that functionally differentiate similar inputs. We show that this negative region plays an active role rather than reflecting a mere gradient optimization side effect. A minimal, sign-specific intervention that zeroes only the negative pre-activations of a small set of Wasserstein neurons substantially increases perplexity and sharply degrades grammatical performance on BLiMP and TSE, whereas both random and perplexity-matched ablations of many more non-Wasserstein neurons in their negative pre-activations leave grammatical performance largely intact. Conversely, on a suite of non-grammatical benchmarks, the perplexity-matched control ablation is more damaging than the Wasserstein neuron ablation, yielding a double dissociation between syntax and other capabilities. Part-of-speech analysis localizes the excess surprisal to syntactic scaffolding tokens, layer-specific interventions show that small local degradations accumulate across depth, and training-dynamics analysis reveals that the same sign-specific ablation becomes more harmful as Wasserstein neurons emerge and stabilize. Together, these results identify negative pre-activations in a sparse subpopulation of Wasserstein neurons as an actively used substrate for syntax in smooth-activation language models.