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.
Learning Symbolic Constraint Representations from Examples: A Neuro-Symbolic Approach
Nassim Belmecheri, Arnaud Gotlieb, Nadjib Lazaar, Helge Spieker
Transformer oracle + FastCA automate human-free constraint acquisition; strong results, but TO3 leaks target scopes and theory assumes a perfect oracle.
Reinforcement Learning over Patient Trajectories for Clinical Reasoning in EHR Foundation Models
Yuxin Xiao, Sheng Zhang, Chandan Singh, Tristan Naumann et al.
Proposes RL fine-tuning (GRPO/DAPO) with time-aware rewards for EHR foundation models; gains are real but modest, single-dataset, with thin statistical rigor.
Decoupling Error Attribution in Cloud-Native Graph-RAG: A Data Integrity Diagnostic Framework
Shuai Yan, Yuhang Wu, Xiaodong Huang, Ke Wang
A three-layer Graph-RAG error audit finds data defects dominate and LLM memory masks them; promising, but tiny single-domain setup and unproven mechanism.
EAR: Entity-Aware Partitioning Approach for Retrieval-Augmented Generation Development
Cenab Batu Bora, Oylum Alatlı, Sebnem Bora, Oguz Dikenelli
EAR cuts retrieved context 37.5-40.2% vs fixed chunks; accuracy effects non-significant on 153 Turkish MCQA items - honest, narrow methodological study.
Breaking the Token Ceiling: Distilling Smaller, Stronger Byte Models
Kalyani Marathe, Artidoro Pagnoni, Tomasz Limisiewicz, Margaret Li et al.
Byte distillation with exact <eot> logit conversion is predicted to beat token distillation by 4% asymptotically; strong study, but wins rest on extrapolation.
Amortized Low-Rank Adaptation for Model-Based Reinforcement Learning
Fernando Palafox, David Fridovich-Keil
CLAW: a hypernetwork emits LoRA adapters for test-time world-model adaptation; well-ablated on small sim families, but narrow scope and no wall-clock numbers.
Robust Prototypical Networks for Few-Shot Sensor Fault Diagnosis
Mohammed Ayalew Belay, Amirshayan Haghipour, Pierluigi Salvo Rossi
MEPN averages prototypes over 10 episodes, lifting one-shot DeFACTO accuracy 73.5%→93.1%, yet equals ProtoNet at K=10 — honest but modest variance reduction.
T-GADE: Thermodynamical Generative-AI-Driven Evolution of LLM Artifacts
Kyoko Ogawa, Naoki Mori
T-GADE adds thermodynamic free-energy selection to LLM heuristic evolution; median training excess drops 29% vs EoH on bin packing; single-task evidence.
Certified AI Triage of ICU Alarms
Mohammed Sameer Syed, Rozhin Yasaei
Claims finite-sample certified suppression of false VT alarms on VTaC; statistically rigorous and unusually transparent, but the bound is marginal and code is withheld.
When Connected Does Not Mean Similar: Charting the Homophily Boundary of SNAP-KG for Streaming Entity Integration
Jui-Chien Lin, Oshani Seneviratne
Companion scope study shows SNAP-KG and homophily-assuming baselines collapse on heterophilous graphs; honest analysis, no new method, roadmap untested.
FRIST: FMRI Representation Informed Shared-space Training Improves EEG-only Individual-Finger BCI Decoding
Jintao Zhang, Yidan Ding, Joshua Kosnoff, Maxim Karrenbach et al.
FRIST uses fMRI-guided training to lift EEG-only finger decoding; gains on 12 participants look solid, but pipeline is complex and offline-only.
Hybrid Physics-AI Framework of Body Center of Mass Dynamics from Wrist-Worn Sensors
Shuhao Que, Valentina Breschi, Ying Wang
Wrist-IMU→COM hybrid physics-AI framework: hybrids reach 5.3–9.3% gait NRMSE; careful analysis, but edge over pure NN not statistically proven, n=10.
Sampling via Decision-Flow: Training-Free Extraction of Improved Latent Reasoning Paths in Large Language Models
Zhendong Mi, Shaoyi Huang
Claims training-free tree sampling beats GRPO/power sampling using GPT-4o terminal scores; gains plausible but judge-conflated and statistically unverified.
Do Influence-Derived Data Perturbations Enable Machine Unlearning? A Controlled Study of Three Plausible Roles
Chenkai Wu, Chrispine Kambimbi, Qinyang Zeng, Jun Yan
Careful role-matched study: influence-derived perturbations (DPL) fail direct unlearning, show inconsistent utility, and add cost as warm start on CIFAR-10.
Self-Verifying Anomaly Detection using Explainable AI for Cybersecurity of DER Networks
Damilola Popoola, Souradeep Bhattacharya, Manimaran Govindarasu
ExCYDER checks LightGBM rules against SHAP to self-verify DER intrusion alerts; promising concept, but verification is unvalidated and evaluation tiny.
AIM: A Privacy-Aware Interoperable Memory Framework for Multi-Agent Multi-User LLM Systems
Zachary Johnson, Nigel Boachie Kumankumah, Somya Chatterjee, Tejas Sathyamurthi et al.
AIM brings index-level privacy control to multi-user LLM memory + MUMBench; promising design but synthetic circularity and no baselines weaken the claims.
Theoretical Guarantees for One-Shot Magnitude Pruning and Compute-Adaptive Early Exit
Erdem Koyuncu
Proves concentration for one-shot pruning and alignment-controlled early-exit tradeoffs; rigorous but idealized, with notable theory-experiment gaps.
ESTS at WMT26: Routing-Informed Expert Pruning for Model Compression
Liu O. Martin, Lucas Bandarkar, Nanyun Peng
Clear, honest WMT26 system paper: GPT-OSS-20B pruned to 4.2–7.8B translation specialists via routing mass; only internal pseudo-reference eval, no ablations.
LoRA-RC: Reservoir Computing with Low-Rank Adaptation
Wenbin Wan
LoRA-RC adapts a reservoir's recurrent matrix online via projected low-rank updates with a sound δISS certificate; but evidence is one Lorenz drift scenario.
ParaRecover: A Process-Level Benchmark for Error Localization and Recovery in Parallel Tool-Use Agents
Bowen Guan, Zhentao Yin, Yanming Shen
Introduces a 10,626-instance process-level benchmark (SDE rubric) for agent error recovery; judge validation solid, but no significance tests or error bars.
I Am No One: Style-Aware Paraphrasing for Text Anonymization
Ahmed Sohair Khan, Estrid He, Monica Wachowicz, Elham Naghizade
LLM style-profile rewriting cuts authorship F1 60-73% with fluency kept; promising, but no significance tests, weak attacker models, missing non-DP baselines.
Toward Robust Personalized Alignment for LLMs: Mitigating Persona Drift in Multi-Turn Dialogue
Youyuan Zhang, Siyuan Li, Fangming Liu, Jing Li
CORE routes persona updates via COMMIT/DEFER/IGNORE to cut drift; broad baselines and ablations support it, but headline gains rest on a GPT-5.1 judge.
SynthSentry: Detecting Synthetic Data Contamination in Language Model Training Data
Praveen Kumar Myakala, Ravichandra Namburi, Sowmya Keragodu Jayaramu, Sooraj George Thomas
Label-free Mahalanobis score screens synthetic corpus contamination; honest but small-scale: recursion untested, composite beaten by one feature.
Inference for Newton Methods with Accelerated Sketch-and-Project via Random Scaling
Xinchen Du, Elizaveta Rebrova, Michał Dereziński, Sen Na
Proves averaged-iterate CLT, FCLT, and pivotal random-scaling inference for accelerated sketched Newton; rigorous theory, shown experiments are synthetic-only.