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.
Exponential Family Synthetic Controls
Hector Rodriguez-Deniz, David M. Blei
Distributional synthetic controls via exponential-family factorization and VI; solid identification theory, but outcome-level gains over SC are marginal.
Rethinking Diffusion Segmentation: When Does It Rely on Its Noisy State, and Does Diffusion Matter?
Hengzhuo Yang, Yuming Zeng, Yuling Yang
Claims diffusion-segmentation state reliance follows supervision path and image-only cores suffice; consistent audit, core split partly theory-guaranteed.
Divergent strategies and convergent outcomes in autonomous materials discovery
Jihan Kim
16 replicated LLM agents diverge in strategy but converge on one frontier—and one database artifact; rigorous pre-registered study, single model-harness only.
UniK: Universal Knowledge Perception for Digital and Physical AI
Nirmit Desai, Kunal Sawarkar, Aditya Mahakali, Dongkon Lee et al.
Claims UniK retrieval + 70B Llama matches frontier LLMs across 5 domains; solid engineering, but frontier baselines lack RAG and leaderboards are self-run.
LEAP-NBV: Lightweight Edge Active-Perception for Foundation-Model Next-Best-View Planning
Boxun Hu, Jiawei Ge, Axel Krieger, Peng Wang et al.
Competent edge-deployment study: distilled HMR + FP16 yields real Jetson gains (3.6 FPS), but evaluation is synthetic-only, with no variance reporting.
Jev-Mem: System-One-Controlled Agentic Memory for Efficient AI Agents
Dongming Jiang, Yi Li, Bingzhe Li
Claims System-One control lifts LoCoMo to 0.777 (+11% rel.) with 6.6x faster builds; promising efficiency design, but single benchmark and no ablations.
The Operational Value of Spatial Dependence in Renewable Forecast Scenarios for Single-Period Economic Dispatch: A Controlled Ablation Study
Jayakumar Manoharan
Claims spatial coherence adds <0.64% dispatch cost in single-period newsvendor dispatch; controlled ablation is rigorous and transparent but scope is narrow.
MobileCybench: Evaluating Agent Vulnerability Discovery via Executable Probes
Andy K. Zhang, Ava Huang, Joey Ji, Wai Han et al.
MobileCybench scores agent-found Android vulnerabilities via executable probes; well-designed and honest, but 13 apps, no significance tests, refusal confounds
ACLArena: Agent Continue Learning in Multi-stage Post-training
Haixin Wang, Xiaoxuan Wang, Junkai Zhang, Han Zhang et al.
ACLArena studies forgetting in sequential agent post-training; strong diagnosis and paradigm comparison, but MLE relies on task-ID routing, 8B-scale only.
AURA: Uncertainty-Routed Activation Editing for Acoustic Grounding in Speech Foundation Models
Natarajan Balaji Shankar, Zilai Wang, Zihan Wang, Mohan Shi et al.
AURA cuts Whisper non-speech hallucination 89.2%→1.9% with ~500x fewer params than LoRA via uncertainty-routed head edits; promising but Whisper-only.
MGRD: Compact morphology-gated residual diffusion for variance-aware cross-domain neurite forecasting
Tsung Yeh Hsieh, Cosmin Anitescu, Chunghwan Kim, Victoria A. Webster-Wood et al.
MGRD: Compact morphology-gated residual diffusion for neurite forecasting; iPSC gains vs gSTA look real, but most stem from the diffusion backbone, not gating.
Simpler Methods Work Better for L1 Penalized Logistic Models and Large Datasets
Edward Raff, James Holt
发现 L1 求解器相关文献仅在极端惩罚下进行测试；旧的 Liblinear 和简单的 LBFGS* 基线在实际 C 值下胜出；但计时混合了停止容差的影响。
RoboTalk: Learning Multi-Robot Communication and Coordination from Multimodal Demonstrations
Dorian Benhamou Goldfajn, Mason Nakamura, Saaduddin Mahmud, Justin Svegliato et al.
Synthetic dual-robot dataset lifts Qwen3-VL-8B from ~2% to 77% held-out success; credible work, but no significance tests and eval mirrors training scaffold.
FinInteract: Benchmarking Clarification and Intent Integration in Ambiguous Financial Question Answering
Xinyu Wang, Tung Sum Thomas Kwok, Zhenghan Tai, Guang Cheng
Frontier LLM agents rarely elicit ambiguity resolutions in financial QA (≤28.9% vs 95% ceiling); strong benchmark and diagnosis, preliminary interventions.
Misaligned Clinical Risk Classification and Cost Asymmetry in Open-Weight Large Language Models
Star S. D. Liu, Xiyu Ding, Robert B. Barrett, Alberto Santamaria-Pang et al.
Claims LLMs encode but fail to integrate clinical cost tradeoffs (2/12 cells cost-correct); rigorous design but one small dataset and no code yet.
ShapeLex: Decoupling Local Shape Symbolization and Global Scale Modeling for Text-Controlled Time Series Generation
Subo Wei, Jianqi Gao, Mingyan Fan, Shaorong Xie et al.
ShapeLex grounds text phrases in discrete shape tokens plus an MDN scale head; benchmark wins are solid but partly condition-driven, and no code is released.
Cost-Accuracy Trade-offs: Neural Operator vs Classical Numerical Solver
Daniel Zhengyu Huang, Andrew M. Stuart
严谨的三基准研究：神经算子仅在低至中等精度下在挂钟时间上优于经典PDE求解器；诚实、可复现、无夸大之处。
Testing, not presuming, adequacy: calibrating generative social simulators against emergent network structure
Tengfei Shao, Chao Li, Xu Wang, Masayuki Goto
Adds SBI-based adequacy testing, repair and held-out audit to social-simulator validation; rigorous and honest, but the tested simulator stays inadequate.
Context-Aware Pre-Deployment Evaluation of AI Systems: A Regulatory Framework for Nigerian Fintech
Andrew Anogie Uduimoh, Hadiza Umar Yusuf, Oluwafemi Osho
Argues Nigerian/AU rules lack pre-deployment AI evaluation specs; a cheap 6-model eval shows large false-positive failures; plausible but statistically thin.
InterHier: Learning Interconnected Hierarchical Semantics for Open-Vocabulary Object Detection
Yeong-Jin Kim, Ho-Joong Kim, Seong-Whan Lee
Learns a prompt context to replace fixed hierarchy connectors in OVD; solid gains over SHiNe, but novelty is thin and OV-COCO gains are unattributed.
When Evidence Conflicts: Reliability-aware Meta-review Generation
Xinzhe Wang, Fei Tao, Jiang Xie, Hong Yu et al.
Reliability-weighted evidence aggregation aids meta-review generation on ORSum (+1.0/+1.2 R-1 vs CGI2); modest gains, single benchmark, LLM-judge circularity.
Q-DEQ: Discrete Solving and Quantization for Deep Equilibrium Models in Time Series Forecasting under Edge Deployment Coding Constraints
Ruotong Yang, Hongdong Zhu, Qi Gao, Yin Ma et al.
Q-DEQ 将 DEQ 的 Anderson 更新重构为 QUBO + W8A8 以用于边缘预测；准确性保持不变（−1.16% 至 +2.90% MSE），但缺乏硬件、运行时或成本证据。
Efficient Reasoning Exploration via State-Conditioned Latent Steering with Progress Guidance
Hengyuan Zhang, Chenming Shang, Zunhai Su, Xiao Liang et al.
SPS: training-free latent steering guiding reasoning toward progress; consistent Pass@4 gains on Qwen3 1.7B-14B across 9 benchmarks, yet stats rigor is thin.
Structured Decomposition for Reliable LLM-Generated Access Control Policies
Vatsal Gupta, Darshan Sreenivasamurthy
Structured pipeline lifts correct Rego generation to 50.3% vs 15.3% (372 ACRE); strong coverage gain, but security claims rest on n≤9 and self-derived tests.