The paper introduces Simplex-Lattice Interpolation Merging (SLIM), a method that builds a quadratic surrogate model of aggregate performance over merging coefficients for large language models using a classical mixture design. By evaluating only individual experts and equal-weight pairs, SLIM identifies the surrogate with the fewest measurements needed and then optimizes it without further target-metric evaluations. Experiments on two model architectures show that SLIM accurately predicts unseen multi-expert mixtures and achieves competitive merge performance under limited evaluation budgets, outperforming random designs and regularized fitting.
By Seongcheol Jeong, Masahiro Suzuki, Yutaka Matsuo
The paper introduces a low-budget active learning approach that selects a small coreset of data points for training high-accuracy models, particularly useful when labeling is expensive, such as in medical contexts. It uses features from a pretrained self-supervised model and applies entropic optimal transport—specifically the Sinkhorn divergence—as the selection criterion, enabling dimension-free sample complexity and efficient gradient-based optimization. The method combines gradient-based candidate generation with a swap-based local search, achieving superior performance over existing heuristics on image and medical datasets.
By Rim Hajal, Mathieu Besan\c{c}on, J\'er\^ome Malick
The paper introduces PPNAS, a prediction‑powered inference method for neural architecture search that combines a small set of accurately evaluated architectures with a large set of zero‑cost proxy (ZCP) evaluations. PPNAS uses the ordinal information from ZCPs to generate pairwise ranking supervision and applies a debiasing step to reconcile discrepancies between proxy and true performance rankings. Experiments show that PPNAS achieves state‑of‑the‑art results in predictor‑based NAS under limited evaluation budgets.
By Pascal Janetzky, Yuxin Wang, Michael Klar, Stefan Feuerriegel
The paper introduces CLEAR, a lightweight, task‑agnostic post‑hoc method that enhances evidential robustness in deep learning models without retraining. CLEAR uses held‑out calibration data to map the geometry of the model’s latent space, then generates perturbation views at inference to detect latent conflict. When high conflict is found, CLEAR selectively reduces evidential strength while preserving evidence for latent‑consistent inputs, achieving significant improvements in OOD and adversarial AUROC on ImageNet→CUB and running much faster than competing methods.
By Charmaine Barker, Daniel Bethell, Simos Gerasimou
Deep Repurposing (DR) is a post‑hoc framework that adapts deep neural networks when parts of their output space become obsolete after deployment. DR estimates the latent geometry of obsolete and retained regions, removes components that support obsolete outputs, and reallocates retained-compatible evidence through an analytic repair map without gradient updates. The method yields repaired predictions and representations that eliminate obsolete outputs while preserving or improving retained accuracy, and it adapts up to 60× faster than competing unlearning methods.
By Daniel Bethell, Charmaine Barker, Simos Gerasimou
MECHVAR is a lightweight, auditable rule for selecting experiments from a finite library to discriminate between candidate mechanisms. It chooses probes by maximizing the posterior‑weighted variance of predicted responses, a score that aligns with the Box–Hill pairwise‑KL criterion and links to expected information gain when separations are small. Experiments on a 25‑block audit and a Digits loop show MECHVAR outperforming confirmation‑first strategies and matching or exceeding EIG in identification accuracy while being far faster to compute.
By Yifan Guo
Faynt is a family of Transformer policies (10M and 75M parameters) that control all 26 characters in Super Smash Bros. Melee from a single checkpoint. After reinforcement learning, the 10M model wins 98.4% of same‑character games against fourteen specialist and multi‑character releases, and defeats a zero‑delay Slippi‑AI model in all 68 evaluated games. The work details architecture, scaling, hyperparameter transfer, supervised pretraining on 840,000 human replays, post‑training curricula, distillation, and efficient inference, and it releases the weights, benchmark suites, and a platform for automated model tournaments.
By Ali Janati, Nikita Kuzmin, Rohit Swamy, Charles Niu
The paper introduces LambdaDockScore, a protein‑protein docking scoring function that leverages the LambdaLoss loss from Learning‑to‑Rank to improve pose ranking. By fine‑tuning the energy prediction head of DFMDock on 2.9 million decoy poses from the DIPS dataset, LambdaDockScore outperforms the state‑of‑the‑art EuDockScore on CAPRI benchmark targets, achieving higher accuracy in top‑1 and top‑5 predictions. It also enhances ranking for antibody‑antigen and protein‑protein complexes with extreme binding interface sizes.
By Richard Zhu, Darren Xu, Lee-Shin Chu, Jeffrey J. Gray
The paper introduces CRAFT, a method for improving compositional generalization in vision‑language‑action (VLA) models. It addresses the issue where models rely on visual shortcuts during fine‑tuning, leading them to execute demonstrated skill combinations that match observations rather than the instructed ones. By training with counterfactual instruction–observation pairs and transferring supervision through reusable skill representations, CRAFT enhances success on unseen skill combinations while preserving performance on demonstrated ones across multiple VLA models and benchmarks.
By Taegeun Yang, Youngju Na, Yoonki Cho, Sung-Eui Yoon
The paper introduces an unsupervised algorithm that separates content and style in language model outputs, enabling explicit style control. Applied to over 100K traces from nine teacher models, the method uncovers six recurring, imbalanced styles and fine‑tunes student models to follow them. This style conditioning improves Pass@$k$ on six math reasoning benchmarks and shows that style influences problem‑solving success, offering a new avenue for controlling and enhancing model performance.
By Ioana Marinescu, Eric Karl Oermann, Kyunghyun Cho
The paper investigates how reinforcement learning agents trained in one cyber simulation can be transferred to other simulators or real environments. It introduces a framework that decouples state alignment from action translation, allowing zero‑shot policy transfer without retraining. Experiments across four cyber platforms show that transferred policies can preserve performance in closely aligned settings and achieve substantial win rates in more divergent environments.
By Sabrina Saika, Yinuo Du, Aritran Piplai
VERITYGATE is a four‑gate framework that checks whether LLM‑generated narrations adhere to a fixed schema of declared evidence IDs, entities, numbers, and claim types, rather than verifying every fact in the prose. In experiments with GPT‑4o‑mini, Llama‑3.3‑70B, and Claude Sonnet 4.6 on 900 grounded‑ungrounded pairs, the framework identified high failure rates (up to 80.3% for mini claims) and demonstrated that a single repair pass can improve claim survival rates. The authors also provide code, data, and preliminary human studies to validate the rules and highlight gaps between schema compliance and correct prose.
"whyItMatters":"The framework offers a systematic way to evaluate and improve the faithfulness of LLM explanations to structured evidence, revealing significant failure rates and the impact of repair strategies."
By Sachin Gupta
TRACE tackles real‑world dynamic resource assignment by combining evolutionary automatic heuristic design with an agentic knowledge‑extraction workflow. A Reasoner agent interprets system logs to hypothesize about underlying dynamics, while a Coder agent generates and runs schema‑specific code to validate these hypotheses, producing insights or executable tools for the evolved heuristics. Evaluations on a synthetic cloud benchmark and a 5G vRAN scenario show that TRACE outperforms existing AHD methods, delivering more auditable heuristics with less than 2% overhead.
By Jose A. Ayala-Romero, Andres Garcia-Saavedra, Xavier Costa-Perez
The paper introduces a framework for aligning the inductive bias of linear time‑invariant State Space Models (SSMs) with task‑specific spectral characteristics. By formalizing the bias through an SSM‑induced kernel and showing its spectrum is governed by the model’s frequency response, the authors propose Task‑Dependent Initialization (TDI), a fast power‑spectrum matching method. Experiments on synthetic data, one‑layer SSMs, and deep SSMs across real‑world benchmarks demonstrate that TDI improves data‑efficient generalization when the task’s spectral structure differs from the default SSM bias.
By Qiyu Chen, Guozhang Chen
JEPA-Bisim introduces a bisimulation encoder to joint-embedding predictive world models, ensuring that states with similar transition dynamics are mapped to nearby latent representations while suppressing irrelevant slow features such as background changes and distractors. The approach improves robustness on navigation (PointMaze) and manipulation (PushT) tasks under varied test-time visual conditions, achieving up to tenfold smaller latent spaces than DINO-WM. It remains effective across different pre-trained visual encoders, including DINOv2, SimDINOv2, and iBOT.
By Leonardo F. Toso, Davit Shadunts, Yunyang Lu, Gloria Geng, Nihal Sharma, Donglin Zhan, Nam H. Nguyen, James Anderson
The paper introduces temporal self‑distillation for reinforcement learning with verifiable rewards (RLVR), proposing that a policy can learn from a stronger future checkpoint of itself. Two methods—Near‑Future Policy Optimization (NPO) and Near‑Future Policy Distillation (NPD)—use verified future‑self trajectories and token‑level transfer, respectively, while AutoNPO adaptively selects the optimal future checkpoint. Experiments on eight image‑text benchmarks show that near‑future teachers yield higher performance than far‑future ones, indicating that the balance between new capability and learner compatibility is key.
By Chuanyu Qin, Chenxu Yang, Qingyi Si, Naibin Gu, Dingyu Yao, Zheng Lin, Peng Fu, Nan Duan, Jiaqi Wang
ConQuR introduces a lightweight post‑training rotation calibration for large language model activation quantization. By learning orthogonal rotations that align normalized activations with the corners of an inscribed hypercube, the method distributes activation energy evenly and can be updated online without storing activations. Experiments on Llama‑2 and Llama‑3 models (3B–70B) show competitive or improved perplexity and reasoning performance while avoiding costly training or large offline storage.
By Chayne Thrash, Ali Abbasi, Soheil Kolouri
The paper proposes constraining the query and key projection matrices in Transformer attention to the Stiefel manifold and optimizing them with a Riemannian Adam optimizer. It demonstrates that this geometric constraint yields significant performance gains on a CIFAR‑10 patch benchmark, with the constrained model outperforming standard AdamW by up to +6.79 percentage points. The authors also show that weight decay has no effect on the constrained frames and that the improvement is driven by a scale‑free step size rather than the manifold projection or equivariance properties.
By Rub\'en Dar\'io Guerrero
The paper introduces a Frequency-aware Decomposition Network (FDN) that estimates vibration-rich wrench signals in sensorless robotic contact tasks. FDN splits the wrench horizon into low-frequency trends and high-frequency residuals, using pointwise regression for the former and a learned conditional distribution for the latter. Experiments on a 6‑DoF hydraulic manipulator show that FDN reduces high‑frequency amplitude error by up to 47% compared to baselines while maintaining low‑frequency accuracy, and can perform 1,000 ms multi‑step‑ahead estimation in 11 ms on a single CPU thread.
By Hyeonbeen Lee, Min-Jae Jung, Tae-Kyeong Yeu, Jong-Boo Han, Daegil Park, Simon Stepputtis, Jin-Gyun Kim
The paper introduces Generalized Biomedicine Discovery (GBD), a benchmark that addresses long‑tail, anomaly, and taxonomy‑aware discovery in medical imaging. It proposes SCAN, a cognition‑inspired method that suppresses expected patterns, highlights surprising deviations, and integrates these into global representations. Experiments demonstrate that SCAN enhances novel concept discovery while preserving established clinical knowledge and can be integrated into existing architectures.
By Luyao Tang, Yingkai Yang, Hanqi Chen, Jiewei Zheng, Chaoqi Chen, Cheng Chen