The study evaluates whether biological reasoning models actually use their biological inputs by testing six models on DNA, protein, and single‑cell tasks. By perturbing one biological input while keeping others fixed, the authors find that many models (e.g., Evo2, ESM3, BioReason, BioReason‑Pro) rely primarily on textual information, with minimal impact from the biological representations. In contrast, models like ChatNT, Prot2Text‑V2, CellWhisperer, and Cell2Sentence‑Scale show greater dependence on their biological inputs, yet overall accuracy gains do not consistently reflect increased biological input contribution.
By Ada Fang, Nikitha Thoduguli, Lukas Fesser, Hanlin Zhang, Sham M. Kakade, Marinka Zitnik
HADRec is a Hierarchy-Aware Drug Recommendation framework that fuses molecular knowledge and electronic health records to improve medication recommendation. It uses LLaMA-7B to encode clinical notes, ChemBERTa to encode drug SMILES strings, and a cross‑attention mechanism for multimodal fusion, while a hierarchical predictor and consistency constraint loss enforce adherence to the ATC classification system. Experiments on MIMIC‑III and MIMIC‑IV show state‑of‑the‑art performance, strong generalization, and well‑calibrated predictions, with counterfactual evaluation indicating clinically aligned reasoning.
By Junke Wang, Hongshun Ling, Li Zhang, Jinjing Wu, Tong Shao, Fang Wang, Yuan Gao
The paper investigates how test‑time computation can enhance language models and at what cost, introducing the SELF‑POT benchmark to evaluate this across competition mathematics, competitive programming, and agentic workflows. SELF‑POT separates candidate coverage from final accuracy, tracks correctness transitions under revision, and measures protocol completion alongside task success. Using a unified budget rule, the study compares direct inference, parallel sampling, and self‑revision across five low‑cost reasoning models, revealing that selection rules and failure handling significantly influence gains and cost savings.
By Bangji Yang, Jingyuan Li, Jiajun Fan, Yi Evie Zhang, Ruihan Guo, Hongba Ma, Neil He, Chumeng Liang, Qinglong Zheng, Zhanghan Ni, Ge Liu
The paper introduces Adaptive Policy-Guided Error Mitigation (APGEM), a context-aware layer that dynamically selects error mitigation strategies—such as ZNE, PEC, CDR, and REM—during quantum reinforcement learning (QRL) training on NISQ devices. APGEM uses policy-level indicators (quantum-state fidelity, policy entropy, cumulative reward, and approximation ratio) to choose the most suitable mitigation method and integrates it directly into the reinforcement learning loop. Evaluated on the Capacitated Vehicle Routing Problem under various NISQ noise models, APGEM outperforms static mitigation techniques, achieving about 94% of an oracle strategy’s utility, maintaining higher fidelity as noise increases, and producing more stable learning behavior.
By Bisma Majid, Shabir Ahmed Sofi, Mir Mohammad Yousuf
The paper introduces Commute-Time-Preserving World Models (CTWMs), which learn latent representations that reflect commute-times in an environment by using a latent displacement predictor and a log-determinant regularizer. This approach addresses the issue that existing self-supervised methods degrade the necessary eigenvalue-dependent scaling for accurate commute-time representation. In experiments, CTWMs outperform the task-agnostic baseline LeWM on several continuous goal-reaching benchmarks while using only half the parameters.
By Michael Hauri, Peter Buttaroni, Fabian A. Mikulasch, Friedemann Zenke
ibUMAP introduces a coherent, field‑based alternative to UMAP’s stochastic negative sampling, evaluating attraction and repulsion from a shared embedding snapshot synchronously. By representing the degree‑weighted repulsive field with three scalar moments and using an interpolation‑based FFT scheme, it avoids explicit all‑pairs computations while maintaining efficient CPU and GPU performance. Benchmarks show median speedups of 3.29× (CPU) and 1.44× (GPU) over existing implementations, with improved run‑to‑run stability and controlled fidelity trade‑offs.
By Bin Chen, Yumeng Xue, Patrick Paetzold, Yunhai Wang, Oliver Deussen
The paper introduces SimpleTimeBench, a diagnostic suite testing basic temporal primitives like monotonic trends, periodic signals, and leading indicator covariates. It finds that prominent multivariate Time Series Foundation Models (Chronos‑2, Moirai, and Toto) often produce suboptimal zero‑shot forecasts for these simple patterns, and that fine‑tuning can improve specific tasks while harming performance on other fundamentals. These failures persist in real‑world sensor forecasting, indicating that current TSFMs may lack the inductive biases needed to capture straightforward relationships, thereby limiting their practical reliability.
By Nafiseh Ghoroghchian, Haipeng Zhang, Shuyi Han, Alex Labach, George Stein
The paper introduces LoopCD, a training‑free contrastive decoding framework that improves token selection in Loop‑Transformer models by comparing the final prediction with earlier recurrent passes. LoopCD operates either in logit space (LoopCD‑Logits) with a single extra output pass or in hidden‑state space (LoopCD‑Hidden) with no output overhead. Across multiple looped Transformer families, LoopCD yields significant performance gains—raising pass@1 scores on tasks such as AIME 2024 and HumanEval—while enabling a reduction in the number of recurrent loops and a corresponding decrease in inference FLOPs.
By Weihao Liu, Huangjie Zheng, Tianrong Chen, Rohit Dilip, Richard He Bai, Yizhu Jiao, Yuyang Wang, Ruixiang Zhang
The paper introduces a supervised end-to-end benchmark for restoring orchestral historical music, addressing the lack of ground-truth pairs in early‑20th‑century recordings. It simulates the historical degradation process more accurately than prior work and trains a latent flow‑matching model that surpasses existing HMR baselines in various evaluations. Additionally, the authors release a 9.3‑hour license‑free test set, code, and audio demos for the community.
By Steven Cho, Junghyun Koo, Raphael Lafargue, Tushar Dhyani, Eloi Moliner, Yuki Mitsufuji
AnchorPrompt is an adaptation technique for large audio‑language models that keeps the base model frozen and learns a single block of prompt vectors inserted at the decoder input. By training these prompts through self‑distillation on diverse audio and text perturbations, the method improves answer consistency and reduces hallucinations across multiple benchmarks. The approach is perturbation‑agnostic at inference, enabling zero‑shot transfer to unseen distortions such as reverberation and choice permutations.
By Pooneh Mousavi, Amir Ivry, Mirco Ravanelli, Cem Subakan
The paper introduces Neural Fourier Surrogates (NFS), a stochastic classical neural network that learns coefficients over a finite Fourier series support similar to quantum neural networks (QNNs). By testing on tabular benchmark datasets, NFS demonstrates competitive classification performance against established classical baselines and data‑reuploading QNNs. The authors also compare the learned Fourier spectra of QNNs and NFS on synthetic data, positioning NFS as a natural classical baseline for evaluating QNN performance.
By Oliver Knitter, Jonathan Mei, Sang Hyub Kim, Chi Chen, Masako Yamada, Martin Roetteler
The paper introduces a degree‑corrected joint matrix factorization technique for detecting communities in multilayer networks. It uses a nonnegative symmetric matrix trifactorization that enforces disjoint, shared communities across layers while allowing each layer to have distinct connectivity patterns and node degrees. An efficient algorithm is presented and evaluated on a multilayer degree‑corrected stochastic block model, showing superior performance compared to existing methods.
By Alexandra Dache, Manon Rustin, Arnaud Vandaele, Nicolas Gillis
The paper introduces ONLINE LLM PICKER, a framework for active model selection of large language models in streaming settings. It selects the most informative prompts for annotation within a limited budget, enabling the identification of the best or near‑best model among many candidates. Experiments on 10 datasets and over 130 language models show up to 71.67% savings in annotation cost and a reduction in regret by up to 2.51× when using the chosen model for sequential generation.
By Alessandro Turrin, Patrik Okanovic, Torsten Hoefler, Nezihe Merve G\"urel
The paper investigates continual reinforcement learning using neuroevolution, comparing evolution strategies (ES) and genetic algorithms (GAs) across diverse environments and network sizes. ES consistently achieves a better balance between stability and plasticity, while GAs are more plastic but forget more. The authors attribute this to ES finding wider neighborhoods in weight space, with overlap between consecutive tasks correlating with the stability-plasticity trade‑off, and note that common RL plasticity issues do not transfer to neuroevolution.
By Eleni Nisioti, Andrea Cossu, Kathrin Korte, Sebastian Risi
The paper introduces Universal Byte-Level Encoding (UBE), a dual‑alphabet tokenizer that routes 3‑4‑byte UTF‑8 characters through UTF‑16 while keeping 1‑2‑byte characters on the UTF‑8 path. This design lowers the worst‑case token‑budget disparity for high‑premium scripts without increasing costs for efficient English spans, and it preserves standard BPE merges and exact decoding. In extensive Unicode audits and multilingual language‑model experiments, UBE matches or improves token‑count efficiency and context usability compared to traditional byte‑pair encoding.
By Hyunsik Kim, Youngmoon Jung
The paper introduces a data‑free method for model merging that estimates per‑layer covariance matrices directly from difference matrices, eliminating the need for auxiliary data. This approach reduces computational costs while maintaining a principled interference‑minimization framework. Experiments on vision and language benchmarks with models from 86 M to 7 B parameters show that the method outperforms existing data‑free merging techniques.
By Marawan Gamal Abdel Hameed, Derek Tam, Pascal Jr Tikeng Notsawo, Colin Raffel, Guillaume Rabusseau
The paper introduces INTRA, an attention-based encoder-decoder framework that retrieves directly from its own internal representations instead of using an external retriever. By having decoder attention query pre-encoded evidence chunks, INTRA unifies retrieval and generation, eliminating the typical mismatch seen in retrieval-augmented generation pipelines. Experiments on question-answering benchmarks show that INTRA outperforms strong engineered retrieval pipelines in both evidence recall and overall answer quality.
By Elad Hoffer, Yochai Blau, Edan Kinderman, Ron Banner, Daniel Soudry, Boris Ginsburg
The paper presents convergence guarantees for Plug-and-Play (PnP) image restoration algorithms that use annealed noise levels, covering both deterministic and stochastic methods. It identifies explicit nonconvex objectives linked to the final denoising level and proves that iterates become asymptotically stationary with respect to these objectives, without requiring a specific noise decay schedule. The authors validate their theory experimentally on tasks such as inpainting, super‑resolution, demosaicing, and tomography.
By Samuel Hurault
The paper audits the IBM Telco Customer Churn benchmark, revealing that common practices inflate performance metrics. It shows that pre‑split SMOTE boosts churn‑class F1 by 13.1 points, that isotonic regression is the best calibration method while temperature scaling fails on tree ensembles, and that the cost‑optimal decision threshold is 5–10 times lower than the F1‑optimal one, saving about $77,000 per 1,000 customers. The authors also test generalisation on Iranian Telecom and Bank churn datasets, and propose a four‑component reporting checklist with reproducible code.
By Soumyadeep Roy
CommunityKV is a new framework that treats sparse attention as a community detection problem, building a token graph from $QK^T$ scores and partitioning it into semantically coherent communities. It updates token communities in constant time during streaming decoding, avoiding costly global re‑partitioning. Experiments on Qwen3 and Llama‑3.1 show that CommunityKV can increase end‑to‑end generation throughput by up to 1.25×, and with query‑group graph aggregation up to 1.71×, while maintaining comparable accuracy.
By Joe McKenna, Anastasios Alexandridis, Nathan Susanj, Jing Liu