Alignment via reinforcement learning from human feedback (RLHF) has become the dominant paradigm for controlling the quality of outputs from large language models (LLMs). However, when viewed as `loss + regularization,' the standard RLHF objective lacks theoretical justification and incentivizes degenerate, deterministic solutions, an issue that variants such as Direct Policy Optimization (DPO) also inherit. In this paper, we rethink alignment by framing it as \emph{distribution learning} from pairwise preference feedback by explicitly modeling how information about the target language model bleeds through the preference data. This explicit modeling leads us to propose three principled learning objectives: preference maximum likelihood estimation, preference distillation, and reverse KL minimization. We theoretically show that all three approaches enjoy strong non-asymptotic $O(1/n)$ convergence to the target language model, naturally avoiding degeneracy and reward overfitting. Finally, we empirically demonstrate that our distribution learning framework, especially preference distillation, consistently outperforms or matches the performances of RLHF and DPO across various tasks and models.

Varying coefficient models are popular for estimating nonlinear regression functions in functional data models. Their Bayesian variants have received limited attention in large data applications, primarily due to prohibitively slow posterior computations using Markov chain Monte Carlo (MCMC) algorithms. We introduce Bayesian data sketching for varying coefficient models to obviate computational challenges presented by large sample sizes. To address the challenges of analyzing large data, we compress the functional response vector and predictor matrix by a random linear transformation to achieve dimension reduction and conduct inference on the compressed data. Our approach distinguishes itself from several existing methods for analyzing large functional data in that it requires neither the development of new models or algorithms, nor any specialized computational hardware while delivering fully model-based Bayesian inference. Well-established methods and algorithms for varying coefficient regression models can be applied to the compressed data. We establish posterior contraction rates for estimating the varying coefficients and predicting the outcome at new locations with the randomly compressed data model. We use simulation experiments and analyze remote sensed vegetation data to empirically illustrate the inferential and computational efficiency of our approach.

Knowledge of accurate relative skills in any competitive system is essential, but foundational approaches such as ELO discard extremely relevant performance data by concentrating exclusively on binary outcomes. While margin of victory (MOV) extensions exist, they often lack a definitive method for incorporating this information. We introduce Margin of Victory Differential Analysis (MOVDA), a framework that enhances traditional rating systems by using the deviation between the true MOV and a $\textit{modeled expectation}$. MOVDA learns a domain-specific, non-linear function (a scaled hyperbolic tangent that captures saturation effects and home advantage) to predict expected MOV based on rating differentials. Crucially, the $\textit{difference}$ between the true and expected MOV provides a subtle and weighted signal for rating updates, highlighting informative deviations in all levels of contests. Extensive experiments on professional NBA basketball data (from 2013 to 2023, with 13,619 games) show that MOVDA significantly outperforms standard ELO and Bayesian baselines. MOVDA reduces Brier score prediction error by $1.54\%$ compared to TrueSkill, increases outcome accuracy by $0.58\%$, and most importantly accelerates rating convergence by $13.5\%$, while maintaining the computational efficiency of the original ELO updates. MOVDA offers a theoretically motivated, empirically superior, and computationally lean approach to integrating performance magnitude into skill rating for competitive environments like the NBA.

Data-carving methods perform selective inference by conditioning the distribution of data on the observed selection event. However, existing data-carving approaches typically require an analytically tractable characterization of the selection event. This paper introduces a new method that leverages tools from flow-based generative modeling to approximate a potentially complex conditional distribution, even when the underlying selection event lacks an analytical description -- take, for example, the data-adaptive tuning of model parameters. The key idea is to learn a transport map that pushes forward a simple reference distribution to the conditional distribution given selection. This map is efficiently learned via a normalizing flow, without imposing any further restrictions on the nature of the selection event. Through extensive numerical experiments on both simulated and real data, we demonstrate that this method enables flexible selective inference by providing: (i) valid p-values and confidence sets for adaptively selected hypotheses and parameters, (ii) a closed-form expression for the conditional density function, enabling likelihood-based and quantile-based inference, and (iii) adjustments for intractable selection steps that can be easily integrated with existing methods designed to account for the tractable steps in a selection procedure involving multiple steps.

Small Uncrewed Aerial Systems (sUAS) are increasingly deployed as autonomous swarms in search-and-rescue and other disaster-response scenarios. In these settings, they use computer vision (CV) to detect objects of interest and autonomously adapt their missions. However, traditional CV systems often struggle to recognize unfamiliar objects in open-world environments or to infer their relevance for mission planning. To address this, we incorporate large language models (LLMs) to reason about detected objects and their implications. While LLMs can offer valuable insights, they are also prone to hallucinations and may produce incorrect, misleading, or unsafe recommendations. To ensure safe and sensible decision-making under uncertainty, high-level decisions must be governed by cognitive guardrails. This article presents the design, simulation, and real-world integration of these guardrails for sUAS swarms in search-and-rescue missions.

Malicious URLs persistently threaten the cybersecurity ecosystem, by either deceiving users into divulging private data or distributing harmful payloads to infiltrate host systems. Gaining timely insights into the current state of this ongoing battle holds significant importance. However, existing reviews exhibit 4 critical gaps: 1) Their reliance on algorithm-centric taxonomies obscures understanding of how detection approaches exploit specific modal information channels; 2) They fail to incorporate pivotal LLM/Transformer-based defenses; 3) No open-source implementations are collected to facilitate benchmarking; 4) Insufficient dataset coverage.This paper presents a comprehensive review of malicious URL detection technologies, systematically analyzing methods from traditional blacklisting to advanced deep learning approaches (e.g. Transformer, GNNs, and LLMs). Unlike prior surveys, we propose a novel modality-based taxonomy that categorizes existing works according to their primary data modalities (URL, HTML, Visual, etc.). This hierarchical classification enables both rigorous technical analysis and clear understanding of multimodal information utilization. Furthermore, to establish a profile of accessible datasets and address the lack of standardized benchmarking (where current studies often lack proper baseline comparisons), we curate and analyze: 1) publicly available datasets (2016-2024), and 2) open-source implementations from published works(2013-2025). Then, we outline essential design principles and architectural frameworks for product-level implementations. The review concludes by examining emerging challenges and proposing actionable directions for future research. We maintain a GitHub repository for ongoing curating datasets and open-source implementations: https://github.com/sevenolu7/Malicious-URL-Detection-Open-Source/tree/master.

Forecasting future events has significant decision-relevance, as having a well-calibrated probabilistic estimation of the risk of a future pandemic, a conflict, or an emerging technology is crucial in making decisions under uncertainty. Current best practices for forecasting rely on aggregating the judgemental forecasts of experienced forecasters (Tetlock & Gardner 2016), a process that is both lengthy and expensive, though it promises to produce valuable input into decision-making processes (Mellers et al, 2019; Tetlock et al. 2014). Recent work has applied frontier large language models (LLM) to forecasting, testing a variety of research questions, such as whether LLMs are able to match human forecasting performance, what determines their prediction capabilities, and how these capabilities may be increased. For example, previous work looked at retrieval-augmented systems (Halawi et al. 2024), aggregation of multiple models (Schoenegger et al. 2024), ranking-based context retrieval systems (Yan et al. 2024), or applications of reinforcement learning (Turtel et al. 2025b). While many of these approaches have resulted in increased forecasting performance, the current performance of frontier models still trails experienced forecaster aggregates on ForecastBench (Karger et al. 2024). Many such approaches have focused on specific aspects in designing forecasting pipelines such as effective news aggregation (Wang et al. 2025) or fine-tuning on model self-play output (Turtel et al. 2025).

Theory-of-Mind (ToM) enables humans to infer mental states-such as beliefs, desires, and intentions-forming the foundation of social cognition. However, existing computational ToM methods rely on structured workflows with ToM-specific priors or deep model fine-tuning, which struggle with scalability in multimodal environments and fail to generalize as task complexity increases. To address these limitations, we propose a scalable Bayesian ToM planner that decomposes ToM reasoning into stepwise Bayesian updates. Our framework introduces weak-to-strong control, allowing smaller language models (LMs) to specialize in ToM-specific likelihood estimation and transfer their reasoning behaviors to larger LMs (7B to 405B) for integration with social and world knowledge. This synergistic approach aligns large-model inference of human mental states with Bayesian principles. Extensive experiments show that our method achieves a 4.6% accuracy improvement over state-of-the-art techniques on multimodal ToM benchmarks, including challenging unseen scenarios, thereby establishing a new standard for modeling human mental states in complex environments.

The Fourier representation for the uniform distribution over the Boolean cube has found numerous applications in algorithms and complexity analysis. Notably, in learning theory, learnability of Disjunctive Normal Form (DNF) under uniform as well as product distributions has been established through such representations. This paper makes five main contributions. First, it introduces a generalized Fourier expansion that can be used with any distribution $D$ through the representation of the distribution as a Bayesian network (BN). Second, it shows that the main algorithmic tools for learning with the Fourier representation, that use membership queries to approximate functions by recovering their heavy Fourier coefficients, can be used with slight modifications with the generalized expansion. These results hold for any distribution. Third, it analyzes the $L_1$ spectral norm of conjunctions under the new expansion, showing that it is bounded for a class of distributions which can be represented by difference bounded tree BN, where a parent node in the BN representation can change the conditional expectation of a child node by at most $α<0.5$. Lower bounds are presented to show that such constraints are necessary. The fourth contribution uses these results to show the learnability of DNF with membership queries under difference bounded tree BN. The final contribution is to develop an algorithm for learning difference-bounded tree BN distributions, thus extending the DNF learnability result to cases where the distribution is not known in advance.

Model reprogramming adapts pretrained models to downstream tasks by modifying only the input and output spaces. Visual reprogramming (VR) is one instance for vision tasks that adds a trainable noise pattern (i.e., a visual prompt) to input images to facilitate downstream classification. The existing VR approaches for CLIP train a single visual prompt using all descriptions of different downstream classes. However, the limited learning capacity may result in (1) a failure to capture diverse aspects of the descriptions (e.g., shape, color, and texture), and (2) a possible bias toward less informative attributes that do not help distinguish between classes. In this paper, we introduce a decoupling-and-reweighting framework. Our decoupled visual prompts (DVP) are optimized using descriptions grouped by explicit causes (DVP-cse) or unsupervised clusters (DVP-cls). Then, we integrate the outputs of these visual prompts with a probabilistic reweighting matrix (PRM) that measures their contributions to each downstream class. Theoretically, DVP lowers the empirical risk bound. Experimentally, DVP outperforms baselines on average across 11 downstream datasets. Notably, the DVP-PRM integration enables insights into how individual visual prompts influence classification decisions, providing a probabilistic framework for understanding reprogramming. Our code is available at https://github.com/tmlr-group/DecoupledVP.