❌

Normal view

Received — 26 March 2026 ⏭

PERMA: Benchmarking Personalized Memory Agents via Event-Driven Preference and Realistic Task Environments

✇cs.AI, q-bio.NC updates on arXiv.org
By: Shuochen Liu · Junyi Zhu · Long Shu · Junda Lin · Yuhao Chen · Haotian Zhang · Chao Zhang · Derong Xu · Jia Li · Bo Tang · Zhiyu Li · Feiyu Xiong · Enhong Chen · Tong Xu
25 March 2026 at 12:00
arXiv:2603.23231v1 Announce Type: new Abstract: Empowering large language models with long-term memory is crucial for building agents that adapt to users' evolving needs. However, prior evaluations typically interleave preference-related dialogues with irrelevant conversations, reducing the task to needle-in-a-haystack retrieval while ignoring relationships between events that drive the evolution of user preferences. Such settings overlook a fundamental characteristic of real-world personalization: preferences emerge gradually and accumulate across interactions within noisy contexts. To bridge this gap, we introduce PERMA, a benchmark designed to evaluate persona consistency over time beyond static preference recall. Additionally, we incorporate (1) text variability and (2) linguistic alignment to simulate erratic user inputs and individual idiolects in real-world data. PERMA consists of temporally ordered interaction events spanning multiple sessions and domains, with preference-related queries inserted over time. We design both multiple-choice and interactive tasks to probe the model's understanding of persona along the interaction timeline. Experiments demonstrate that by linking related interactions, advanced memory systems can extract more precise preferences and reduce token consumption, outperforming traditional semantic retrieval of raw dialogues. Nevertheless, they still struggle to maintain a coherent persona across temporal depth and cross-domain interference, highlighting the need for more robust personalized memory management in agents. Our code and data are open-sourced at https://github.com/PolarisLiu1/PERMA.

ImplicitRM: Unbiased Reward Modeling from Implicit Preference Data for LLM alignment

✇cs.AI, q-bio.NC updates on arXiv.org
By: Hao Wang · Haocheng Yang · Licheng Pan · Lei Shen · Xiaoxi Li · Yinuo Wang · Zhichao Chen · Yuan Lu · Haoxuan Li · Zhouchen Lin
25 March 2026 at 12:00
arXiv:2603.23184v1 Announce Type: cross Abstract: Reward modeling represents a long-standing challenge in reinforcement learning from human feedback (RLHF) for aligning language models. Current reward modeling is heavily contingent upon experimental feedback data with high collection costs. In this work, we study \textit{implicit reward modeling} -- learning reward models from implicit human feedback (e.g., clicks and copies) -- as a cost-effective alternative. We identify two fundamental challenges in implicit reward modeling: (1) Implicit preference data lacks definitive negative samples, which makes standard positive-negative classification methods inapplicable; (2) Implicit preference data suffers from user preference bias, where different responses have different propensities to elicit user feedback actions, which exacerbates the difficulty of distinguishing definitive negative samples. To address these challenges, we propose ImplicitRM, which aims to learn unbiased reward models from implicit preference data. ImplicitRM stratifies training samples into four latent groups via a stratification model. Building on this, it derives a learning objective through likelihood maximization, which we prove is theoretically unbiased, effectively resolving both challenges. Experiments demonstrate that ImplicitRM learns accurate reward models across implicit preference datasets. Code is available on our project website.

TRACE: A Multi-Agent System for Autonomous Physical Reasoning in Seismological

✇cs.AI, q-bio.NC updates on arXiv.org
By: Feng Liu · Jian Xu · Xin Cui · Xinghao Wang · Zijie Guo · Jiong Wang · S. Mostafa Mousavi · Xinyu Gu · Hao Chen · Ben Fei · Lihua Fang · Fenghua Ling · Zefeng Li · Lei Bai
25 March 2026 at 12:00
arXiv:2603.21152v2 Announce Type: replace-cross Abstract: Inferring the physical mechanisms that govern earthquake sequences from indirect geophysical observations remains difficult, particularly across tectonically distinct environments where similar seismic patterns can reflect different underlying processes. Current interpretations rely heavily on the expert synthesis of catalogs, spatiotemporal statistics, and candidate physical models, limiting reproducibility and the systematic transfer of insight across settings. Here we present TRACE (Trans-perspective Reasoning and Automated Comprehensive Evaluator), a multi-agent system that combines large language model planning with formal seismological constraints to derive auditable, physically grounded mechanistic inference from raw observations. Applied to the 2019 Ridgecrest sequence, TRACE autonomously identifies stress-perturbation-induced delayed triggering, resolving the cascading interaction between the Mw 6.4 and Mw 7.1 mainshocks; in the Santorini-Kolumbo case, the system identifies a structurally guided intrusion model, distinguishing fault-channeled episodic migration from the continuous propagation expected in homogeneous crustal failure. By providing a generalizable logical infrastructure for interpreting heterogeneous seismic phenomena, TRACE advances the field from expert-dependent analysis toward knowledge-guided autonomous discovery in Earth sciences.

Deciphering lung adenocarcinoma heterogeneity: a multi-omics approach reveals nuclear division fibroblasts as prognosticators and therapeutic targets

✇Omics In Lung
By: Peng Cao · Yongxia Qin · Dingtao Hu · Tengfei Ge · Peng Zhang · Chao Cheng · Shun Wang · Shuhui Hu · Zhen Zhang · Ling Xu
21 March 2026 at 18:00

J Transl Med. 2026 Mar 20. doi: 10.1186/s12967-026-08022-3. Online ahead of print.

ABSTRACT

BACKGROUND: Lung adenocarcinoma (LUAD) is a predominant contributor to cancer‑related mortality globally. Lung‑associated fibroblasts (LAFs) are intricately linked to tumorigenesis and the tumor microenvironment (TME), but their heterogeneity and prognostic relevance in LUAD remain incompletely understood. This study aimed to systematically characterize LAF subsets across the spectrum of pulmonary disease, identify LAF subpopulations associated with LUAD prognosis, and construct a robust LAF‑based prognostic signature.

METHODS: We employed a multi-omics approach, leveraging bulk RNA data of 2719 patients from 19 LUAD cohorts, single-cell RNA (scRNA) sequencing data of 368,904 cells from 93 samples, and spatial transcriptomics data of 15,673 spots from 6 samples to characterize the landscape of LAFs across various stages of pulmonary disease. We employed multiple advanced machine learning algorithms to construct and validate a robust nuclear division LAFs (nLAFs) risk score (nLRS) prediction model.

RESULTS: We observed a dynamic and gradual increase in the proportion of LAFs during the progression of LUAD. Throughout this process, we identified nine LAFs subtypes and found nLAFs are significantly associated with the prognosis of LUAD. Utilizing 100 machine learning algorithm combinations and integrating nLAFs marker genes, we developed a five gene based nLRS model, which demonstrated superior performance than other 49 published models in predicting clinical outcomes for LUAD. Additionally, we observed distinct biological functions and immune cell infiltration in the TME between high and low nLRS groups. Exploratory analysis of pan-cancer immunotherapy cohorts suggested that patients with high nLRS scores may exhibit resistance to immunotherapy in some cancer types, but prospective validation in LUAD-specific cohorts is required. Conversely, high nLRS patients displayed increased sensitivity to chemotherapeutic and targeted therapies in preclinical models.

CONCLUSION: Our study introduces a candidate five-gene signature derived from nLAFs that may serve as a robust prognostic biomarker pending prospective validation, offering insights into personalized therapeutic strategies for LUAD patients.

PMID:41862916 | DOI:10.1186/s12967-026-08022-3

Received — 17 March 2026 ⏭

One-Step Flow Policy: Self-Distillation for Fast Visuomotor Policies

✇cs.AI, q-bio.NC updates on arXiv.org
By: Shaolong Li · Lichao Sun · Yongchao Chen
16 March 2026 at 12:00
arXiv:2603.12480v1 Announce Type: cross Abstract: Generative flow and diffusion models provide the continuous, multimodal action distributions needed for high-precision robotic policies. However, their reliance on iterative sampling introduces severe inference latency, degrading control frequency and harming performance in time-sensitive manipulation. To address this problem, we propose the One-Step Flow Policy (OFP), a from-scratch self-distillation framework for high-fidelity, single-step action generation without a pre-trained teacher. OFP unifies a self-consistency loss to enforce coherent transport across time intervals, and a self-guided regularization to sharpen predictions toward high-density expert modes. In addition, a warm-start mechanism leverages temporal action correlations to minimize the generative transport distance. Evaluations across 56 diverse simulated manipulation tasks demonstrate that a one-step OFP achieves state-of-the-art results, outperforming 100-step diffusion and flow policies while accelerating action generation by over $100\times$. We further integrate OFP into the $\pi_{0.5}$ model on RoboTwin 2.0, where one-step OFP surpasses the original 10-step policy. These results establish OFP as a practical, scalable solution for highly accurate and low-latency robot control.

BitDance: Scaling Autoregressive Generative Models with Binary Tokens

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yuang Ai · Jiaming Han · Shaobin Zhuang · Weijia Mao · Xuefeng Hu · Ziyan Yang · Zhenheng Yang · Yali Wang · Huaibo Huang · Xiangyu Yue · Hao Chen
16 March 2026 at 12:00
arXiv:2602.14041v2 Announce Type: replace-cross Abstract: We present BitDance, a scalable autoregressive (AR) image generator that predicts binary visual tokens instead of codebook indices. With high-entropy binary latents, BitDance lets each token represent up to $2^{256}$ states, yielding a compact yet highly expressive discrete representation. Sampling from such a huge token space is difficult with standard classification. To resolve this, BitDance uses a binary diffusion head: instead of predicting an index with softmax, it employs continuous-space diffusion to generate the binary tokens. Furthermore, we propose next-patch diffusion, a new decoding method that predicts multiple tokens in parallel with high accuracy, greatly speeding up inference. On ImageNet 256x256, BitDance achieves an FID of 1.24, the best among AR models. With next-patch diffusion, BitDance beats state-of-the-art parallel AR models that use 1.4B parameters, while using 5.4x fewer parameters (260M) and achieving 8.7x speedup. For text-to-image generation, BitDance trains on large-scale multimodal tokens and generates high-resolution, photorealistic images efficiently, showing strong performance and favorable scaling. When generating 1024x1024 images, BitDance achieves a speedup of over 30x compared to prior AR models. We release code and models to facilitate further research on AR foundation models. Code and models are available at: https://github.com/shallowdream204/BitDance.
Received — 14 March 2026 ⏭

FNDC1 Competitively Binds Gbeta2 to Suppress the beta-Catenin-Destruction Complex and Promote Gastric Cancer Malignancy

✇Omics in Gastric
By: Wenyu Gao · Hao Chen · Fangyu Lin · Jialin Liu · Tingxuan Huang · Han Wang · Xiaojiao Weng · Xinli Wang · Xiaoyan Lin · Tao Jiang
11 March 2026 at 18:00

FASEB J. 2026 Mar 31;40(6):e71634. doi: 10.1096/fj.202503587R.

ABSTRACT

Gastric cancer (GC) is a leading cause of cancer-related deaths and has high recurrence rate. Although fibronectin domain-containing protein 1 (FNDC1) is implicated in GC progression, its molecular mechanisms remain unclear. Multi-omics analyses (TCGA, GEO datasets) were used to assess FNDC1 expression and clinical correlation. In vitro (cell proliferation, invasion, EMT markers) and in vivo (xenograft) experiments, combined with molecular assays (Co-IP, WB, ChIP), explored FNDC1's function and mechanism. FNDC1 was significantly upregulated in GC, correlating with advanced clinicopathological features and poor prognosis. Knockdown of FNDC1 suppressed GC cell proliferation, invasion, and metastasis by inhibiting EMT and Wnt/β-catenin signaling. Mechanistically, FNDC1 competitively bound the WD5 domain (residues 224-254) of Gβ2, disrupting Gβγ-Dvl1 interaction. This prevented Dvl1 degradation, promoted Axin1 ubiquitination, and destabilized the β-catenin-destruction complex (GSK3 β-APC-Axin1), leading to β-catenin accumulation and Wnt pathway activation. FNDC1 drives GC malignancy by targeting the Gβ2-Dvl1 axis to activate Wnt/β-catenin signaling, suggesting FNDC1 as a novel prognostic biomarker and therapeutic target.

PMID:41808415 | PMC:PMC12976582 | DOI:10.1096/fj.202503587R

Received — 11 March 2026 ⏭

Deconstructing Multimodal Mathematical Reasoning: Towards a Unified Perception-Alignment-Reasoning Paradigm

✇cs.AI, q-bio.NC updates on arXiv.org
By: Tianyu Yang · Sihong Wu · Yilun Zhao · Zhenwen Liang · Lisen Dai · Chen Zhao · Minhao Cheng · Arman Cohan · Xiangliang Zhang
10 March 2026 at 12:00
arXiv:2603.08291v1 Announce Type: new Abstract: Multimodal Mathematical Reasoning (MMR) has recently attracted increasing attention for its capability to solve mathematical problems that involve both textual and visual modalities. However, current models still face significant challenges in real-world visual math tasks. They often misinterpret diagrams, fail to align mathematical symbols with visual evidence, and produce inconsistent reasoning steps. Moreover, existing evaluations mainly focus on checking final answers rather than verifying the correctness or executability of each intermediate step. To address these limitations, a growing body of recent research addresses these issues by integrating structured perception, explicit alignment, and verifiable reasoning within unified frameworks. To establish a clear roadmap for understanding and comparing different MMR approaches, we systematically study them around four fundamental questions: (1) What to extract from multimodal inputs, (2) How to represent and align textual and visual information, (3) How to perform the reasoning, and (4) How to evaluate the correctness of the overall reasoning process. Finally, we discuss open challenges and offer perspectives on promising directions for future research.

\$OneMillion-Bench: How Far are Language Agents from Human Experts?

✇cs.AI, q-bio.NC updates on arXiv.org
By: Qianyu Yang · Yang Liu · Jiaqi Li · Jun Bai · Hao Chen · Kaiyuan Chen · Tiliang Duan · Jiayun Dong · Xiaobo Hu · Zixia Jia · Yang Liu · Tao Peng · Yixin Ren · Ran Tian · Zaiyuan Wang · Yanglihong Xiao · Gang Yao · Lingyue Yin · Ge Zhang · Chun Zhang · Jianpeng Jiao · Zilong Zheng · Yuan Gong
10 March 2026 at 12:00
arXiv:2603.07980v1 Announce Type: cross Abstract: As language models (LMs) evolve from chat assistants to long-horizon agents capable of multi-step reasoning and tool use, existing benchmarks remain largely confined to structured or exam-style tasks that fall short of real-world professional demands. To this end, we introduce \$OneMillion-Bench \$OneMillion-Bench, a benchmark of 400 expert-curated tasks spanning Law, Finance, Industry, Healthcare, and Natural Science, built to evaluate agents across economically consequential scenarios. Unlike prior work, the benchmark requires retrieving authoritative sources, resolving conflicting evidence, applying domain-specific rules, and making constraint decisions, where correctness depends as much on the reasoning process as the final answer. We adopt a rubric-based evaluation protocol scoring factual accuracy, logical coherence, practical feasibility, and professional compliance, focused on expert-level problems to ensure meaningful differentiation across agents. Together, \$OneMillion-Bench provides a unified testbed for assessing agentic reliability, professional depth, and practical readiness in domain-intensive scenarios.

Oscillatory shear stress-driven endothelial-to-mesenchymal transition: a critical mechanical signal transduction mechanism in atherosclerosis progression

✇Cell Death Discovery nature.com science feeds
By: Jie Li · Wenchao Xu · Jie Ju · Min Liu · Wenxu Wang · Min Cheng · Xiaoyun Zhang · Xiaodong Cui · Hao Chen
10 March 2026 at 08:00

Cell Death Discovery, Published online: 10 March 2026; doi:10.1038/s41420-026-03000-6

Oscillatory shear stress-driven endothelial-to-mesenchymal transition: a critical mechanical signal transduction mechanism in atherosclerosis progression
Received — 5 March 2026 ⏭

LifeBench: A Benchmark for Long-Horizon Multi-Source Memory

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zihao Cheng · Weixin Wang · Yu Zhao · Ziyang Ren · Jiaxuan Chen · Ruiyang Xu · Shuai Huang · Yang Chen · Guowei Li · Mengshi Wang · Yi Xie · Ren Zhu · Zeren Jiang · Keda Lu · Yihong Li · Xiaoliang Wang · Liwei Liu · Cam-Tu Nguyen
5 March 2026 at 13:00
arXiv:2603.03781v1 Announce Type: new Abstract: Long-term memory is fundamental for personalized agents capable of accumulating knowledge, reasoning over user experiences, and adapting across time. However, existing memory benchmarks primarily target declarative memory, specifically semantic and episodic types, where all information is explicitly presented in dialogues. In contrast, real-world actions are also governed by non-declarative memory, including habitual and procedural types, and need to be inferred from diverse digital traces. To bridge this gap, we introduce Lifebench, which features densely connected, long-horizon event simulation. It pushes AI agents beyond simple recall, requiring the integration of declarative and non-declarative memory reasoning across diverse and temporally extended contexts. Building such a benchmark presents two key challenges: ensuring data quality and scalability. We maintain data quality by employing real-world priors, including anonymized social surveys, map APIs, and holiday-integrated calendars, thus enforcing fidelity, diversity and behavioral rationality within the dataset. Towards scalability, we draw inspiration from cognitive science and structure events according to their partonomic hierarchy; enabling efficient parallel generation while maintaining global coherence. Performance results show that top-tier, state-of-the-art memory systems reach just 55.2\% accuracy, highlighting the inherent difficulty of long-horizon retrieval and multi-source integration within our proposed benchmark. The dataset and data synthesis code are available at https://github.com/1754955896/LifeBench.

EvoPrune: Early-Stage Visual Token Pruning for Efficient MLLMs

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yuhao Chen · Bin Shan · Xin Ye · Cheng Chen
5 March 2026 at 13:00
arXiv:2603.03681v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) have shown strong performance in vision-language tasks, but their inference efficiency is severely limited by the exponential growth of visual tokens in complex scenarios such as high-resolution images and videos. Existing visual token pruning methods mainly operate after visual encoding, overlooking the substantial computational cost incurred during the encoding stage. To address this issue, we propose EvoPrune, an early-stage visual token pruning method for MLLMs that performs pruning directly during visual encoding. Specifically, EvoPrune employs a layer-wise pruning strategy guided by token similarity, diversity, and attention-based importance to retain the most informative visual tokens at selected encoding layers. Extensive experiments on image and video benchmarks validate the effectiveness of EvoPrune. In particular, on the VideoMME dataset, EvoPrune achieves 2$\times$ inference speedup with less than 1% performance degradation, demonstrating its potential for latency-sensitive MLLM deployment.

R1-Code-Interpreter: LLMs Reason with Code via Supervised and Multi-stage Reinforcement Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yongchao Chen · Yueying Liu · Junwei Zhou · Yilun Hao · Jingquan Wang · Yang Zhang · Na Li · Chuchu Fan
5 March 2026 at 13:00
arXiv:2505.21668v3 Announce Type: replace Abstract: Practical guidance on training Large Language Models (LLMs) to leverage Code Interpreter across diverse tasks remains lacking. We present R1-Code-Interpreter, an extension of a text-only LLM trained via multi-turn supervised fine-tuning (SFT) and reinforcement learning (RL) to autonomously generate multiple code queries during step-by-step reasoning. Unlike prior RL + tool-use efforts focused on narrow domains such as math or retrieval, we curate 144 diverse reasoning and planning tasks and show that training a general-purpose Code Interpreter across them presents significant challenges due to task heterogeneity and scarcity of effective samples. To address this, we introduce a multi-stage curriculum learning approach that partitions training samples by measured improvement potential. The RL training prioritizes samples with higher potential and gradually shifts to lower-potential ones, increasing the average RL gains from merely +3.4% to +9.3% across Qwen-2.5 models (3/7/14B). Our final model, R1-CI-14B, improves average accuracy on the 37 test tasks from 44.1% to 72.4%, outperforming text-only GPT-4o (58.6%) and GPT-4o with Code Interpreter (70.9%). Notably, R1-CI-14B also exhibits emergent self-checking behavior through code generation. Datasets, Codes, and Models are available at https://github.com/yongchao98/R1-Code-Interpreter and https://huggingface.co/yongchao98.

Can a Small Model Learn to Look Before It Leaps? Dynamic Learning and Proactive Correction for Hallucination Detection

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zepeng Bao · Shen Zhou · Qiankun Pi · Jianhao Chen · Mayi Xu · Ming Zhong · Yuanyuan Zhu · Tieyun Qian
5 March 2026 at 13:00
arXiv:2511.05854v2 Announce Type: replace Abstract: Hallucination in large language models (LLMs) remains a critical barrier to their safe deployment. For hallucination detection to be practical in real-world scenarios, the use of efficient small models is essential to ensure low latency and minimal resource consumption. However, existing methods rely on fixed verification strategies, where simply tuning small models to mimic fixed verification trajectories fails to capture the adaptability required for diverse hallucination patterns, thereby inducing planning instability. To address this limitation, we propose a ``Learning to Evaluate and Adaptively Plan'' (LEAP) framework, which shifts hallucination detection from fixed execution to dynamic strategy learning. Specifically, LEAP first employs a powerful teacher model to iteratively explore and refine verification strategies through a failure-driven loop. This dynamic planning capability is then distilled into an efficient student model, augmented by a novel proactive correction mechanism that enables the model to evaluate and optimize its verification strategy before execution. Experiments on three benchmarks demonstrate that LEAP outperforms state-of-the-art methods, offering an effective and scalable solution for reliable hallucination detection.

Uni-NTFM: A Unified Foundation Model for EEG Signal Representation Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhisheng Chen · Yingwei Zhang · Qizhen Lan · Tianyu Liu · Huacan Wang · Yi Ding · Ziyu Jia · Ronghao Chen · Kun Wang · Xinliang Zhou
5 March 2026 at 13:00
arXiv:2509.24222v2 Announce Type: replace-cross Abstract: Current foundation models for electroencephalography (EEG) rely on architectures adapted from computer vision or natural language processing, typically treating neural signals as pixel grids or token sequences. This approach overlooks that the neural activity is activated by diverse sparse coding across a complex geometric topological cortex. Inspired by biological neural mechanisms, we propose the Unified Neural Topological Foundation Model (Uni-NTFM), an architecture rooted in three core neuroscience principles. In detail, to align with the brain's decoupled coding mechanism, we design the Heterogeneous Feature Projection Module. This module simultaneously encodes both time-domain non-stationary transients and frequency-domain steady-state rhythms, ensuring high quality in both waveform morphology and spectral rhythms. Moreover, we introduce a Topological Embedding mechanism to inject structured spatial priors and align different sensor configurations onto a unified latent functional topography, effectively reconstructing the geometry of brain regions. Furthermore, we achieve functional modularization and sparse coding efficiency of biological networks by constructing the Mixture-of-Experts Transformer network. This dynamic routing mechanism assigns different signal patterns and tasks to specialized neural subnetworks, and effectively preventing task interference while increasing the model capacity to record-breaking 1.9 billion parameters. Uni-NTFM is pre-trained on a diverse corpus comprising 28,000 hours of EEG data, and outperforms existing models across nine distinct downstream tasks under both linear probing and fine-tuning settings, demonstrating that aligning model architecture with neural mechanisms is significant to learn universal representations and achieve generalizable brain decoding. Our code is available at: https://anonymous.4open.science/r/Uni-NTFM-0924.

Efficient Self-Evaluation for Diffusion Language Models via Sequence Regeneration

✇cs.AI, q-bio.NC updates on arXiv.org
By: Linhao Zhong · Linyu Wu · Wen Wang · Yuling Xi · Chenchen Jing · Jiaheng Zhang · Hao Chen · Chunhua Shen
4 March 2026 at 13:00
arXiv:2603.02760v1 Announce Type: cross Abstract: Diffusion large language models (dLLMs) have recently attracted significant attention for their ability to enhance diversity, controllability, and parallelism. However, their non-sequential, bidirectionally masked generation makes quality assessment difficult, underscoring the need for effective self-evaluation. In this work, we propose DiSE, a simple yet effective self-evaluation confidence quantification method for dLLMs. DiSE quantifies confidence by computing the probability of regenerating the tokens in the entire generated sequence, given the full context. This method enables more efficient and reliable quality assessment by leveraging token regeneration probabilities, facilitating both likelihood estimation and robust uncertainty quantification. Building upon DiSE, we further introduce a flexible-length generation framework, which adaptively controls the sequence length based on the model's self-assessment of its own output. We analyze and validate the feasibility of DiSE from the perspective of dLLM generalization, and empirically demonstrate that DiSE is positively correlated with both semantic coherence and answer accuracy. Extensive experiments on likelihood evaluation, uncertainty quantification, and flexible-length generation further confirm the effectiveness of the proposed DiSE.

Enhancing Generative Auto-bidding with Offline Reward Evaluation and Policy Search

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhiyu Mou · Yiqin Lv · Miao Xu · Qi Wang · Yixiu Mao · Jinghao Chen · Qichen Ye · Chao Li · Rongquan Bai · Chuan Yu · Jian Xu · Bo Zheng
4 March 2026 at 13:00
arXiv:2509.15927v4 Announce Type: replace-cross Abstract: Auto-bidding is a critical tool for advertisers to improve advertising performance. Recent progress has demonstrated that AI-Generated Bidding (AIGB), which learns a conditional generative planner from offline data, achieves superior performance compared to typical offline reinforcement learning (RL)-based auto-bidding methods. However, existing AIGB methods still face a performance bottleneck due to their inherent inability to explore beyond the static dataset with feedback. To address this, we propose \textbf{AIGB-Pearl} (\emph{\textbf{P}lanning with \textbf{E}valu\textbf{A}tor via \textbf{RL}}), a novel method that integrates generative planning and policy optimization. The core of AIGB-Pearl lies in constructing a trajectory evaluator to assess the quality of generated scores and designing a provably sound KL-Lipschitz-constrained score-maximization scheme to ensure safe and efficient exploration beyond the offline dataset. A practical algorithm that incorporates the synchronous coupling technique is further developed to ensure the model regularity required by the proposed scheme. Extensive experiments on both simulated and real-world advertising systems demonstrate the state-of-the-art performance of our approach.
Received — 25 February 2026 ⏭

Does Your Reasoning Model Implicitly Know When to Stop Thinking?

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zixuan Huang · Xin Xia · Yuxi Ren · Jianbin Zheng · Xuanda Wang · Zhixia Zhang · Hongyan Xie · Songshi Liang · Zehao Chen · Xuefeng Xiao · Fuzhen Zhuang · Jianxin Li · Yikun Ban · Deqing Wang
24 February 2026 at 13:00
arXiv:2602.08354v2 Announce Type: replace Abstract: Recent advancements in large reasoning models (LRMs) have greatly improved their capabilities on complex reasoning tasks through Long Chains of Thought (CoTs). However, this approach often results in substantial redundancy, impairing computational efficiency and causing significant delays in real-time applications. Recent studies show that longer reasoning chains are frequently uncorrelated with correctness and can even be detrimental to accuracy. In a further in-depth analysis of this phenomenon, we surprisingly uncover and empirically verify that LRMs implicitly know the appropriate time to stop thinking, while this capability is obscured by current sampling paradigms. Motivated by this, we introduce SAGE (Self-Aware Guided Efficient Reasoning), a novel sampling paradigm that unleashes this efficient reasoning potential. Furthermore, integrating SAGE as mixed sampling into group-based reinforcement learning (SAGE-RL) enables SAGE-RL to effectively incorporate SAGE-discovered efficient reasoning patterns into standard pass@1 inference, markedly enhancing both the reasoning accuracy and efficiency of LRMs across multiple challenging mathematical benchmarks.

STCast: Adaptive Boundary Alignment for Global and Regional Weather Forecasting

✇cs.AI, q-bio.NC updates on arXiv.org
By: Hao Chen · Tao Han · Jie Zhang · Song Guo · Lei Bai
24 February 2026 at 13:00
arXiv:2509.25210v2 Announce Type: replace-cross Abstract: To gain finer regional forecasts, many works have explored the regional integration from the global atmosphere, e.g., by solving boundary equations in physics-based methods or cropping regions from global forecasts in data-driven methods. However, the effectiveness of these methods is often constrained by static and imprecise regional boundaries, resulting in poor generalization ability. To address this issue, we propose Spatial-Temporal Weather Forecasting (STCast), a novel AI-driven framework for adaptive regional boundary optimization and dynamic monthly forecast allocation. Specifically, our approach employs a Spatial-Aligned Attention (SAA) mechanism, which aligns global and regional spatial distributions to initialize boundaries and adaptively refines them based on attention-derived alignment patterns. Furthermore, we design a Temporal Mixture-of-Experts (TMoE) module, where atmospheric variables from distinct months are dynamically routed to specialized experts using a discrete Gaussian distribution, enhancing the model's ability to capture temporal patterns. Beyond global and regional forecasting, we evaluate our STCast on extreme event prediction and ensemble forecasting. Experimental results demonstrate consistent superiority over state-of-the-art methods across all four tasks.
Received — 17 February 2026 ⏭

Experiential Reinforcement Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Taiwei Shi · Sihao Chen · Bowen Jiang · Linxin Song · Longqi Yang · Jieyu Zhao
17 February 2026 at 13:00
arXiv:2602.13949v1 Announce Type: cross Abstract: Reinforcement learning has become the central approach for language models (LMs) to learn from environmental reward or feedback. In practice, the environmental feedback is usually sparse and delayed. Learning from such signals is challenging, as LMs must implicitly infer how observed failures should translate into behavioral changes for future iterations. We introduce Experiential Reinforcement Learning (ERL), a training paradigm that embeds an explicit experience-reflection-consolidation loop into the reinforcement learning process. Given a task, the model generates an initial attempt, receives environmental feedback, and produces a reflection that guides a refined second attempt, whose success is reinforced and internalized into the base policy. This process converts feedback into structured behavioral revision, improving exploration and stabilizing optimization while preserving gains at deployment without additional inference cost. Across sparse-reward control environments and agentic reasoning benchmarks, ERL consistently improves learning efficiency and final performance over strong reinforcement learning baselines, achieving gains of up to +81% in complex multi-step environments and up to +11% in tool-using reasoning tasks. These results suggest that integrating explicit self-reflection into policy training provides a practical mechanism for transforming feedback into durable behavioral improvement.
❌