AdaptiveCDM is a modular framework for source‑free few‑shot domain adaptation in cell detection, enabling a pretrained model to adapt to new imaging domains using only a handful of labeled target images and no source data. It combines Resolution‑Aware Augmentation (RAug) to balance scarce, class‑imbalanced samples while preserving cellular morphology, and Category‑Aware Representation Learning (CARL) to strengthen class‑consistent proposals for better localization and classification. Experiments on M5 and Raabin‑WBC datasets show that AdaptiveCDM achieves competitive or superior mAP scores compared to state‑of‑the‑art methods under their respective supervision settings.
By Nimra Dilawar, Sara Nadeem, Javed Iqbal, Waqas Sultani, Mohsen Ali
The CXR‑LT 2026 Challenge introduces a multi‑center, long‑tailed chest X‑ray classification benchmark with over 145,000 radiologist‑annotated images from PadChest and NIH datasets. It defines two core tasks: robust multi‑label classification on 30 known classes and open‑world generalization to 6 unseen rare disease classes. The paper outlines data collection, annotation, solution strategies, and evaluates performance across head‑vs‑tail, calibration, and cross‑center gaps, noting that vision‑language models improve in‑distribution and zero‑shot performance but rare‑finding detection under multi‑center shift remains difficult.
By Hexin Dong, Yi Lin, Pengyu Zhou, Fengnian Zhao, Alan Clint Legasto, Juno Cho, Dohui Kim, Justin Namuk Kim, Mingeon Kim, Sunwoo Kwak, Gabriel Moy\`a-Alcover, Ky Trung Nguyen, Thanh-Huy Nguyen, Ha-Hieu Pham, Huy-Hieu Pham, Huy Le Pham, Nikhileswara Rao Sulake, Aina Tur-Serrano, Ruichi Zhang, Ang Zu, Adam E. Flanders, Zhiyong Lu, Ronald M. Summers, Mingquan Lin, Hao Chen, Yuzhe Yang, George Shih, Yifan Peng
SurgMotion is a video-native foundation model that replaces pixel-level reconstruction with latent motion prediction for surgical video analysis. It introduces motion-guided masked prediction, spatiotemporal affinity self-distillation, and spatiotemporal feature diversity regularization to focus on semantically meaningful regions and avoid representation collapse. Trained on SurgMotion-15M, the largest surgical video dataset, it outperforms state-of-the-art methods across 17 benchmarks, improving workflow recognition, action triplet recognition, skill assessment, polyp segmentation, and depth estimation.
By Jinlin Wu, Felix Holm, Chuxi Chen, An Wang, Yaxin Hu, Xiaofan Ye, Zelin Zang, Miao Xu, Lihua Zhou, Huai Liao, Danny T. M. Chan, Ming Feng, Wai S. Poon, Hongliang Ren, Dong Yi, Nassir Navab, Gaofeng Meng, Jiebo Luo, Hongbin Liu, Zhen Lei
The paper investigates selective on‑policy distillation, where a student model is trained only on token positions chosen by a selector. It demonstrates that the commonly used shared learning rate is not neutral: performance varies significantly with the learning rate for different selectors, leading to inconsistent comparisons. The authors attribute this selector‑rate entanglement to the selection process itself and recommend reporting the full arm‑by‑rate matrix for fair evaluation.
By Chencheng Zhu
StationPDE is a station-oriented surface PDE learning model designed for multi-station multivariate weather forecasting. It builds a terrain-aware continuous surface field from discrete station observations and separates its physical evolution into surface wind transport and upper-air inference, the latter approximating missing upper-air effects via learnable horizontal diffusion. The model also includes a data-driven diffusion branch and an adaptive router to combine forecasts, achieving a 9.6% average MSE reduction over state-of-the-art baselines on Weather2K and MeteoNet datasets.
By Xiao Wang, Changjian Chen, Rongwen Li, Hongwu Liu, Kun Fang, Zhuo Tang
DDGAD introduces a novel approach to graph anomaly detection that focuses on the disagreement between node-wise and contextual estimates rather than on their combined state. By adapting the Adapt-Then-Combine framework, DDGAD generates separate node-wise and neighborhood-dependent estimates, accumulating their pre-consensus disagreement across iterations to identify anomalies. The method is theoretically grounded with graph-spectral and source-response analyses, and empirical results on six benchmarks demonstrate superior AUROC performance compared to existing techniques.
By Yuxin Yang, Limei Hu, Feng Chen
The paper introduces Unsupervised Graph Collective Anomaly Detection (UGCAD), a framework that uses a variational graph autoencoder to learn graph representations of IoT network traffic and then enhances clustering to group nodes. UGCAD identifies collective anomalies by aggregating normal clusters and applying anomaly scores to the refined groups. Experiments on CICIoT2023 and ToN-IoT datasets show that UGCAD outperforms traditional and state‑of‑the‑art clustering‑based CAD methods in both clustering quality and anomaly detection accuracy.
By Dalila Khettaf, Djamel Djenouri, Zeinab Rezaeifar, Youcef Djenouri
CleanScore is a black‑box audit framework that uses only scored outputs to assess whether models have been exposed to benchmark questions. It creates a public form and two fresh, independently written forms for each question, reports an interval for the public‑form advantage, and employs a private negative‑control bank with a transport radius to separate exposure from normal form mismatch. In a registered audit of five open models on 200 GSM8K and 200 ARC‑Challenge items, CleanScore found no exposure‑consistent advantage, bounding surface‑form inflation below five points, while also demonstrating how leaked items can inflate accuracy on unseen paraphrases and how planted advantages can be partially detected even after rewriting.
whyItMatters":"The study shows that CleanScore can detect and quantify exposure effects in benchmark models, providing a more nuanced understanding of model performance beyond simple accuracy scores."
By Jeffery Opoku, David Banahene
The paper investigates Hierarchical Reasoning Models (HRM), a class of hierarchical Transformer-based latent reasoning models, across Sudoku, Maze, and ARC-AGI-2 tasks. By comparing HRM to Transformer baselines, applying causal interventions on recurrent states, and conducting linear probe and sparse autoencoder ablations, the authors find that recurrent models outperform one-pass baselines, that high- and low-level states contribute differently across tasks, and that ablations of sparse autoencoder features cause larger behavioral changes than probe-direction ablations. The study concludes that HRM implements constraint‑aware iterative refinement on a puzzle‑specific solution state, with component contributions varying without a compact, causally important feature set.
By Leo Raphael Rodrigues, Jian Kang
TWIG (Time‑Causal Wavelet Operator for Irregular Graphs) is a graph‑native neural operator designed for autoregressive surrogate modeling on static irregular graphs. It transforms each node’s history into causal multiscale temporal features, separating recent changes from slower memory components, and propagates these through graph‑wavelet operator blocks with gated pointwise channel mixing. The architecture is causal by construction, enabling closed‑loop forecasting where predictions are recursively reused as future inputs, and it consistently outperforms non‑time‑causal baselines across three irregular‑domain forecasting problems.
By Subashree Venkatasubramanian, David A. Barajas-Solano, Chuyang Liu, Daniel M. Tartakovsky, Dipankar Dwivedi
The paper evaluates coreset selection methods by incorporating both selection and training time into a unified wall‑clock budget, using a standardized benchmark across four datasets and multiple selectors. Across numerous budget anchors, simple random or full‑data training consistently outperforms sophisticated selectors, and selection costs are dominated by a full‑dataset scan that cannot be amortized. The study also identifies when subset reuse can justify selection and reports several correctness fixes in a popular codebase.
By Yangze Liu, Zhongyi Han
Merge++ is a post‑hoc refinement technique for model merging that synthesizes task‑representative images by inverting expert checkpoints and then distills expert knowledge into a merged model. It operates without any additional data beyond the checkpoints and can be applied universally across existing weight‑space merging algorithms. Experiments show consistent improvements, with average gains of +2 to +8 points and up to +25.9 on specific configurations.
By Aditya Pola, Vineeth N. Balasubramanian
The paper introduces CASP-LLM, a coverage‑aware semantic prompting framework that mitigates bias toward dominant temporal patterns in prompt‑based time series forecasting. Unlike traditional similarity‑based retrieval that selects top‑K candidates by cosine similarity, CASP‑LLM uses usage‑tracking and a saturating‑gate regularizer to diversify prompt selection without adding learnable parameters. Experiments on six long‑term and the M4 short‑term benchmarks show that CASP‑LLM matches or outperforms similarity‑based LLM forecasters in most settings, with failures traced to cross‑batch usage rather than within‑retrieval redundancy.
By Daeun Ji, Minkyoung Kim, Dongkuk Kim, Yohan Lee, Beomsoo Kim, Beakcheol Jang
The paper introduces the first controlled benchmark for optimizers in discrete diffusion models, evaluating seven optimizers (AdamW, Lion, Muon, SOAP, MARS, MARS‑M, Schedule‑Free) across four diffusion formulations: masked diffusion on text8, uniform diffusion on QM9 and LM1B, and Gaussian diffusion on CelebA‑64. Each optimizer undergoes the same search protocol and is retrained with full budget and multiple seeds, revealing that AdamW, while strong, is not universally optimal and that optimizers validated on autoregressive language models (Muon, MARS‑M, SOAP) can outperform tuned AdamW on certain tasks.
By Arman Bolatov, Egor Shulgin, David Li, Abduragim Shtanchaev, Sebastian U. Stich, Maxim Panov, Eric Moulines, Peter Richt\'arik, Martin Tak\'a\v{c}
The paper benchmarks six long‑tail loss functions—cross‑entropy, weighted CE, class‑balanced loss, focal loss, LDAM, and logit‑adjusted softmax—across three single‑cell foundation model architectures (scGPT, scBERT, Geneformer) and three datasets (Multiple Sclerosis, Zheng68K, human Pancreas). It shows that overall accuracy masks systematic failures on rare, disease‑relevant cell types, with a consistent gap between overall accuracy, Macro‑F1, and rare‑class recall under plain cross‑entropy. The study identifies two distinct regimes of rare‑class failure, predicts reweighting efficacy by absolute training‑set size, and finds class‑balanced loss and LDAM to be the most reliable across all settings.
By Zeyu Dong, Jiahui Zhong
The paper introduces CNA, an AI‑oriented comprehensive normalized assessment that transforms vital‑sign distributions into a standard normal space to produce a unified health‑status score. This score serves as a criterion for evaluating and terminating reinforcement learning policies in Renal Replacement Therapy (RRT). Using a 23‑dimensional state representation and matrix decomposition to handle missing data, the authors apply offline RL algorithms and show that the learned strategy reduces mortality from 13.2% to 5.0% and shortens hospital stays by nearly 59 hours compared to physician‑observed treatments.
By Jiang Liu, Chan Zhou, Yujie Li, Di Wu, Yihao Xie, Peiwei Li, Xin Shu, Jiaqi Zhu, Chunyong Yang, Yuwen Chen, Bin Yi
The paper proposes a hybrid PID–Deep Reinforcement Learning (DRL) controller for industrial processes, addressing the limitations of traditional PID controllers in complex, non‑linear, multi‑input environments. Using the Industrial Benchmark (IB) to test DRL, the authors develop a multi‑objective reward function and employ a TD3 agent to discover optimal settings for the IB’s ‘Gain’ and ‘Shift’ parameters. These parameters are then fed into a tuned PID controller, yielding a system that combines the optimal performance and efficiency of DRL with the reliability of classical control.
By Zhengyang (Cissy), Gu, Joseph E. Hernandez, John Burtenshaw, Sean Scott, Thomas Cook, Chris Couch
The study compares six state‑of‑the‑art sleep‑staging foundation models on the MESA polysomnography dataset under three signal conditions: EEG only, ECG only, and EEG+ECG. Results show that EEG alone yields the highest accuracy (BIOT macro‑F1 = 0.7237), while switching to ECG alone incurs a consistent accuracy loss of about 0.35 macro‑F1 and reduces data rate to one‑third. Adding ECG to EEG offers little benefit for most models, indicating that EEG carries most of the sleep‑staging signal and that wearable‑compatible ECG alone is a viable but less accurate alternative.
By Hassan Mehdi, Riku Klen, Ayse Kosal Bulbul, Suzanne Timmons, Abdulhamit Subasi, Wei Chen, Zou Zhu, Muhammad Irfan
The study investigates whether language models can explicitly report constraints they have learned through post‑training fine‑tuning. Using constrained recipe generation with five banned ingredients, the authors compare supervised fine‑tuning (SFT) and Group Relative Policy Optimization (GRPO) against an untrained baseline on a Constraint Awareness Benchmark. Both fine‑tuning methods increase behavioral compliance from 4% to about 90% but reduce explicit constraint reporting and erode retained third‑person knowledge, with GRPO showing more destructive effects. The results suggest that reward‑based signals may suppress constraints context‑independently, and that models fail to enumerate constraints on request even when they can avoid them internally.
By Arin Agarwal
JAREX is a Bayesian active‑learning acquisition function designed for multi‑objective process characterization in pharmaceutical manufacturing. It treats characterization as a joint boundary‑learning problem, adaptively selecting experiments to delineate the joint pass region where multiple quality thresholds are simultaneously met. Benchmarks show that JAREX outperforms factorial design of experiments, space‑filling designs, and greedy strategies, achieving more accurate and sample‑efficient boundary recovery, and halving the number of experiments needed for batched runs while maintaining high accuracy.
By Xinyang Li, Kevin Stone, Ajit Vikram