Recent work shows activation steering substantially improves language models' Theory of Mind (ToM) (Bortoletto et al. 2024), yet the mechanisms of what changes occur internally that leads to different outputs remains unclear. We propose decomposing ToM in LLMs by comparing steered versus baseline LLMs' activations using linear probes trained on 45 cognitive actions. We applied Contrastive Activation Addition (CAA) steering to Gemma-3-4B and evaluated it on 1,000 BigToM forward belief scenarios (Gandhi et al. 2023), we find improved performance on belief attribution tasks (32.5\% to 46.7\% accuracy) is mediated by activations processing emotional content : emotion perception (+2.23), emotion valuing (+2.20), while suppressing analytical processes: questioning (-0.78), convergent thinking (-1.59). This suggests that successful ToM abilities in LLMs are mediated by emotional understanding, not analytical reasoning.

Two-stage Stochastic Programming (2SP) is a standard framework for modeling decision-making problems under uncertainty. While numerous methods exist, solving such problems with many scenarios remains challenging. Selecting representative scenarios is a practical method for accelerating solutions. However, current approaches typically rely on clustering or Monte Carlo sampling, failing to integrate scenario information deeply and overlooking the significant impact of the scenario order on solving time. To address these issues, we develop HGCN2SP, a novel model with a hierarchical graph designed for 2SP problems, encoding each scenario and modeling their relationships hierarchically. The model is trained in a reinforcement learning paradigm to utilize the feedback of the solver. The policy network is equipped with a hierarchical graph convolutional network for feature encoding and an attention-based decoder for scenario selection in proper order. Evaluation of two classic 2SP problems demonstrates that HGCN2SP provides high-quality decisions in a short computational time. Furthermore, HGCN2SP exhibits remarkable generalization capabilities in handling large-scale instances, even with a substantial number of variables or scenarios that were unseen during the training phase.

Large-scale genomic workflows used in precision medicine can process datasets spanning tens to hundreds of gigabytes per sample, leading to high memory spikes, intensive disk I/O, and task failures due to out-of-memory errors. Simple static resource allocation methods struggle to handle the variability in per-chromosome RAM demands, resulting in poor resource utilization and long runtimes. In this work, we propose multiple mechanisms for adaptive, RAM-efficient par-allelization of chromosome-level bioinformatics workflows. First, we develop a symbolic regression model that estimates per-chromosome memory consumption for a given task and introduces an interpolating bias to conservatively minimize over-allocation. Second, we present a dynamic scheduler that adaptively predicts RAM usage with a polynomial regression model, treating task packing as a Knapsack problem to optimally batch jobs based on predicted memory requirements. Additionally, we present a static scheduler that optimizes chromosome processing order to minimize peak memory while preserving throughput. Our proposed methods, evaluated on simulations and real-world genomic pipelines, provide new mechanisms to reduce memory overruns and balance load across threads. We thereby achieve faster end-to-end execution, showcasing the potential to optimize large-scale genomic workflows.

Accurate 6D object pose estimation is vital for robotics, augmented reality, and scene understanding. For seen objects, high accuracy is often attainable via per-object fine-tuning but generalizing to unseen objects remains a challenge. To address this problem, past arts assume access to CAD models at test time and typically follow a multi-stage pipeline to estimate poses: detect and segment the object, propose an initial pose, and then refine it. Under occlusion, however, the early-stage of such pipelines are prone to errors, which can propagate through the sequential processing, and consequently degrade the performance. To remedy this shortcoming, we propose four novel extensions to model-based 6D pose estimation methods: (i) a dynamic non-uniform dense sampling strategy that focuses computation on visible regions, reducing occlusion-induced errors; (ii) a multi-hypothesis inference mechanism that retains several confidence-ranked pose candidates, mitigating brittle single-path failures; (iii) iterative refinement to progressively improve pose accuracy; and (iv) series of occlusion-focused training augmentations that strengthen robustness and generalization. Furthermore, we propose a new weighted by visibility metric for evaluation under occlusion to minimize the bias in the existing protocols. Via extensive empirical evaluations, we show that our proposed approach achieves more than 5% improvement in accuracy on ICBIN and more than 2% on BOP dataset benchmarks, while achieving approximately 3 times faster inference.

While small language models (SLMs) have shown promise on various reasoning tasks, their ability to judge the correctness of answers remains unclear compared to large language models (LLMs). Prior work on LLM-as-a-judge frameworks typically relies on comparing candidate answers against ground-truth labels or other candidate answers using predefined metrics like entailment. However, this approach is inherently indirect and difficult to fully automate, offering limited support for fine-grained and scalable evaluation of reasoning outputs. In this work, we propose JudgeBoard, a novel evaluation pipeline that directly queries models to assess the correctness of candidate answers without requiring extra answer comparisons. We focus on two core reasoning domains: mathematical reasoning and science/commonsense reasoning, and construct task-specific evaluation leaderboards using both accuracy-based ranking and an Elo-based rating system across five benchmark datasets, enabling consistent model comparison as judges rather than comparators. To improve judgment performance in lightweight models, we propose MAJ (Multi-Agent Judging), a novel multi-agent evaluation framework that leverages multiple interacting SLMs with distinct reasoning profiles to approximate LLM-level judgment accuracy through collaborative deliberation. Experimental results reveal a significant performance gap between SLMs and LLMs in isolated judging tasks. However, our MAJ framework substantially improves the reliability and consistency of SLMs. On the MATH dataset, MAJ using smaller-sized models as backbones performs comparatively well or even better than their larger-sized counterparts. Our findings highlight that multi-agent SLM systems can potentially match or exceed LLM performance in judgment tasks, with implications for scalable and efficient assessment.

Structured pruning removes entire neurons or channels, but its effectiveness depends on how importance is distributed across the representation space. Change-of-basis (CoB) pruning addresses this challenge by applying orthogonal linear transformations that concentrate importance within certain dimensions. However, many standard deep learning architectures are not inherently invariant to such transformations. To enable compatibility, we introduce two-subspace radial activations (TSRAs): an activation family that is invariant to orthogonal linear transformations applied independently within its two activation subspaces. This invariance allows CoB transformations to be merged into surrounding weights without incurring extra parameters. We position this work as a proof-of-concept that a rotationally invariant design may offer a principled approach towards change-of-basis pruning. We do not provide an analysis of multiple TSRA candidates nor do we explore weight initialization for any TSRAs. These limitations, combined with other necessary modifications we make to permit rotational invariance, result in a slight accuracy drop of $4.52\%$ compared to a ReLU-based control. However, using activation-magnitude importance, VGG-16 implementing our CoB+TSRA framework shows encouraging results on CIFAR-10. Under fixed-ratio structured pruning, CoB improves accuracy over a TSRA baseline at all pruning ratios and extends reliable pruning frontier from roughly $30\%$ to $70\%$ of parameters without post-prune fine tuning. Under threshold-based pruning strategies, CoB prunes $90-96\%$ of parameters while maintaining $1-6\%$ accuracy drop after fine-tuning. Together, these results indicate that rotationally invariant architectures may offer a promising path towards CoB pruning.

The performance of machine learning models in drug discovery is highly dependent on the quality and consistency of the underlying training data. Due to limitations in dataset sizes, many models are trained by aggregating bioactivity data from diverse sources, including public databases such as ChEMBL. However, this approach often introduces significant noise due to variability in experimental protocols. We introduce AssayMatch, a framework for data selection that builds smaller, more homogenous training sets attuned to the test set of interest. AssayMatch leverages data attribution methods to quantify the contribution of each training assay to model performance. These attribution scores are used to finetune language embeddings of text-based assay descriptions to capture not just semantic similarity, but also the compatibility between assays. Unlike existing data attribution methods, our approach enables data selection for a test set with unknown labels, mirroring real-world drug discovery campaigns where the activities of candidate molecules are not known in advance. At test time, embeddings finetuned with AssayMatch are used to rank all available training data. We demonstrate that models trained on data selected by AssayMatch are able to surpass the performance of the model trained on the complete dataset, highlighting its ability to effectively filter out harmful or noisy experiments. We perform experiments on two common machine learning architectures and see increased prediction capability over a strong language-only baseline for 9/12 model-target pairs. AssayMatch provides a data-driven mechanism to curate higher-quality datasets, reducing noise from incompatible experiments and improving the predictive power and data efficiency of models for drug discovery. AssayMatch is available at https://github.com/Ozymandias314/AssayMatch.

Large Language Model (LLM) Agents, often trained with Reinforcement Learning (RL), are constrained by a dependency on human-curated data, limiting scalability and tethering AI to human knowledge. Existing self-evolution frameworks offer an alternative but are typically restricted by the model's inherent capabilities and single-round interactions, hindering the development of complex curricula involving tool use or dynamic reasoning. We introduce Agent0, a fully autonomous framework that evolves high-performing agents without external data through multi-step co-evolution and seamless tool integration. Agent0 establishes a symbiotic competition between two agents initialized from the same base LLM: a curriculum agent that proposes increasingly challenging frontier tasks, and an executor agent that learns to solve them. We integrate external tools to enhance the executor's problem-solving capacity; this improvement, in turn, pressures the curriculum agent to construct more complex, tool-aware tasks. Through this iterative process, Agent0 establishes a self-reinforcing cycle that continuously produces high-quality curricula. Empirically, Agent0 substantially boosts reasoning capabilities, improving the Qwen3-8B-Base model by 18% on mathematical reasoning and 24% on general reasoning benchmarks. Code is available at https://github.com/aiming-lab/Agent0.

Digital agriculture is transforming the way we grow food by utilizing technology to make farming more efficient, sustainable, and productive. This modern approach to agriculture generates a wealth of valuable data that could help address global food challenges, but farmers are hesitant to share it due to privacy concerns. This limits the extent to which researchers can learn from this data to inform improvements in farming. This paper presents the Digital Agriculture Sandbox, a secure online platform that solves this problem. The platform enables farmers (with limited technical resources) and researchers to collaborate on analyzing farm data without exposing private information. We employ specialized techniques such as federated learning, differential privacy, and data analysis methods to safeguard the data while maintaining its utility for research purposes. The system enables farmers to identify similar farmers in a simplified manner without needing extensive technical knowledge or access to computational resources. Similarly, it enables researchers to learn from the data and build helpful tools without the sensitive information ever leaving the farmer's system. This creates a safe space where farmers feel comfortable sharing data, allowing researchers to make important discoveries. Our platform helps bridge the gap between maintaining farm data privacy and utilizing that data to address critical food and farming challenges worldwide.

Abstract--Many robots are not equipped with a manipulator and many objects are not suitable for prehensile manipulation (such as large boxes and cylinders). In these cases, pushing is a simple yet effective non-prehensile skill for robots to interact with and further change the environment. Existing work often assumes a set of predefined pushing modes and fixed-shape objects. This work tackles the general problem of controlling a robotic fleet to push collaboratively numerous arbitrary objects to respective destinations, within complex environments of cluttered and movable obstacles. It incorporates several characteristic challenges for multi-robot systems such as online task coordination under large uncertainties of cost and duration, and for contact-rich tasks such as hybrid switching among different contact modes, and under-actuation due to constrained contact forces. The proposed method is based on combinatorial hybrid optimization over dynamic task assignments and hybrid execution via sequences of pushing modes and associated forces. It consists of three main components: (I) the decomposition, ordering and rolling assignment of pushing subtasks to robot subgroups; (II) the keyframe guided hybrid search to optimize the sequence of parameterized pushing modes for each subtask; (III) the hybrid control to execute these modes and transit among them. Last but not least, a diffusion-based accelerator is adopted to predict the keyframes and pushing modes that should be prioritized during hybrid search; and further improve planning efficiency. The framework is complete under mild assumptions. Its efficiency and effectiveness under different numbers of robots and general-shaped objects are validated extensively in simulations and hardware experiments, as well as generalizations to heterogeneous robots, planar assembly and 6D pushing. Humans often interact with objects via non-prehensile skills such as pushing and rolling, especially when prehensile skills such as stable grasping is infeasible. This aspect is however less exploited in robotic systems. Most existing work treats pushing as a complementary skill to pick-and-place primitives for a single manipulator within simple environments, e.g., [1], [2], [3], [4]. Nonetheless, pushing can be particularly beneficial for low-cost mobile robots that are not equipped with a manipulator, e.g., ground vehicles, quadruped robots, and even underwater vehicles [5]. For instance, obstacles can be pushed out of the path, and target objects can be pushed to desired positions.