Evolve Quick Reviews
Public Paper Reviews
Structured preliminary reviews of recent open-access papers, generated and continuously updated by the Evolve analysis engine. Each review condenses what the paper claims, how solid the evidence is, and the questions a rigorous peer reviewer would ask — English analysis with abstracts in your language. These are independent AI analyses, not peer review; the canonical record always lives at the publisher.
GraphToolbox: A Configurable Python Framework for Graph Neural Network Forecasting
Eloi Campagne, Yvenn Amara-Ouali, Yannig Goude, Argyris Kalogeratos
GraphToolbox unifies GNN load forecasting (51 PyG convs); French-load gains are real but modest, classical baselines win on net-load. Honest, rigorous software.
UK-PRBENCH: A Paragraph-Level Precedent Retrieval Benchmark for United Kingdom Case Law
Damith Premasiri, Tharindu Ranasinghe
First paragraph-level UK precedent retrieval benchmark from 64k judgments; valuable, but recall-only metrics and unexplained fine-tuning collapses weaken it.
Written as a Record, Read as an Address: What a Forward Pass Leaves in an Operation's KV Cache
Lingfeng Wu, Behzad Shomali
Claims LLM forward passes write causally recoverable KV-cache records the model itself reads only as addresses; sound causal design, mostly synthetic tasks.
Custom Named Entity Recognition and Topic Classification for Global Health Publications
Genis Skura, Antoine Geissbühler, Jean-Luc Falcone
Geneva master's thesis builds a spaCy NER + topic-classification pipeline for global health; practical and candid, but evaluations use single-annotator data.
Augmented Hypothesis Testing with Persona-Based LLM Simulations
Ziyad Benomar, Aymen Al Marjani, Paul Missault, Saab Mansour
Asymmetric Z-test + GPPI (transformed PPI++) leverage LLM-persona predictions in A/B testing with solid theory; four-dataset gains real but often modest.
Ascent: An Agentic System over the Model Context Protocol for Real-World Clinical Data Analysis
Angelo Ziletti, Leonardo D'Ambrosi, Melanie Tuchardt, Tim Kondziella
MCP-based clinical agent + EpiTrap error benchmark: capable models gain +27/+20 pp over fixed pipeline, but no compute-matched baseline and judge bias remain.
Learning tactile perception from high-bandwidth single-point sensing
Joseph Rigal, Emmanuel Virot, Caroline Pascal
SpectRobot turns single-point vibration into spectrogram images for ACT policies; occluded sorting up to ~92% with long windows; candid but small-n study.
DUMA-Bench: A Dual-Control Multi-Agent Benchmark for Evaluating LLM Agent Security
Ivan Aleksandrov, German Kochnev, Sabrina Sadiekh, Yaroslav Rogoza
Bench 增加了对抗性双控制（基于 τ²-bench）；ASR 在14个模型上从 26.9% 跃升至 41.1%；设计合理但 ASR 混淆了失败情况且统计数据未报告。
From Semantic Decisions to Feasible Trajectories: Self-Evolving LLM-Guided Optimal Control for Narrow-Space Parking
Zhengbao Yao, Yuanfu Luo, Kehan Xue
LLM decides maneuvers, OCP enforces feasibility for narrow-space parking; promising but small 38-scene simulation, no code, evolution protocol unclear.
Annie, Are You Okay? How Style- and Context-Based Personalization Shape AI-Assisted Decision-Making
Hasibur Rahman, Benjamin R. Cowan, Smit Desai
Claims context-based personalization sways stock choices with no trust change; preregistered N=240 experiment is well controlled, with minor hypothetical-stakes caveats.
Qwen-Audio-3.1-Realtime: Towards Reliable Agentic Voice Interaction
Lujia Bao, Qian Chen, Luyao Cheng, Chong Deng et al.
Qwen-Audio-3.1-Realtime reports broad agentic voice gains via distillation and GRPO; extensive results, but no ablations, CIs, or independent verification.
Circuit Hypernetworks for Quantum-Augmented Diffusion Language Models
Xiaoqiang Wang, Mengyang Xiong, Jun Dai, Bang Liu
HyperQ emits per-token IQP circuits in a frozen 1.1B diffusion LM; width gains look real but rest on three widths, no error bars, unverified baselines.
iSDFT: Information-Proximal Self-Distillation for Continual Learning in LLMs
Ahmed Khaled Khamis, Xiaotong Ji, Hassan Jaber, Rasul Tutunov et al.
Budgeted teacher information via I-projection improves on-policy self-distillation on 4 LLM backbones; single-seed evidence and capped math evals temper gains.
Assessing Readability with LLMs: The Role of Reasoning and Few-Shot Prompting
Raphaël Thieffry, Matej Martinc
Careful benchmark of 5 open LLMs on readability levels: native reasoning and 1-shot prompting help most; findings credible but lacks significance testing.
Corrective Forcing: Unified Post-Training for Diffusions and Flows in Generative Speech Enhancement
Qing Yao, Lijian Gao, Qirong Mao
CoF post-training curbs exposure bias in diffusion/flow speech enhancement with NFE-robust gains (PESQ 3.07→3.21), but still trails predictive baselines.
Touch2Robot: Robot Touch in the Human Demonstration Loop
Shengcheng Luo, Xiaoyang Cheng, Hong Ying, Xiaoying Zhou et al.
Simulated robot-touch feedback in human demos nearly doubles replay success (37.9→72.1%) and lifts downstream policy; strong real-robot study, one embodiment.
Trust in Edge-Enabled IoT Security: Features, Challenges and Research Directions
Esin Ece Aydın, Şerif Bahtiyar, Gürkan Gür
Survey of 55 studies builds a lifecycle×domain taxonomy of edge-IoT trust management; timely and structured, but search bias and ad-hoc ratings limit rigor.
TimeLitmus: A Diagnostic Benchmark for Cross-Modal Understanding and Explanation Faithfulness in Event-Conditioned Time-Series Prediction
Jie Gong, Maowei Jiang, Zhiwei Liu, Yankai Chen et al.
Solid diagnostic benchmark: LLMs' event-conditioned time-series accuracy overstates cross-modal skill; caveat: rule-based labels, small pairs.
Guaranteed Low-Rank Tensor Recovery from Modewise Measurements via Normalized Block-Weighted Riemannian Gradient Descent
Yushi Zhou, Feng Zhang
Block-weighted modewise RGD for Tucker recovery: linear convergence proven, worst-case bound no better than unweighted; gains empirical, synthetic only.
Muon Can Outperform Dedicated Continual Learning Methods
Sebastian George Sincari, Bogdan Alexandru Gheorghe, Antonio Barbalau
IncLoRA+Muon matches O-LoRA/ELLA; Muon spreads updates over ~7 singular dirs vs AdamW's 1.4–1.8. Small-scale (T5) but honest and mechanistically convincing.
Understanding Hyperspherical Geometry of ECAPA-TDNN Embedding and Its Impact on Zero-Shot Voice Conversion
Mathilde Abrassart, Nicolas Obin, Axel Roebel
Claims geometric regularization of ECAPA prototypes improves zero-shot VC; coherent evidence but single-seed, no term ablation - promising preliminary study.
Adapting Tree-Structured Speculative Decoding to DeepSeek-V4 for Efficient Inference
Changxu Liu, Zhaogeng Li
Credible systems adaptation of tree speculation to DeepSeek-V4; acceptance gains consistent, throughput up to 18.5%, but evaluation narrow and no code release.
Multi-Term Fourier Graph Neural Network with Sample Relationship Learning for Enhanced Remaining Useful Life Prediction
Ya Song, Laurens Bliek, Yaoxin Wu, Yingqian Zhang
Fourier GNN + multi-window sampling + inter-sample graph learning for RUL; CMAPSS SOTA claimed (avg RMSE 11.52), but single-benchmark eval, no seed variance.
Reasoning Topology Matters: A Controlled Study of LLM-Based Cybersecurity Analysis
Jiling Zhou, Aisvarya Adeseye, Antti Hakkala, Seppo Virtanen et al.
GoT prompting beats CoT/few-shot by 9.8–12.2pp on 3 security tasks; trend credible, but Table II's perfectly uniform +4.0pp gaps undermine trust.