Liu et al. develop Oncoformer, a multimodal transformer that reads routine laboratory tests and chest X-rays already collected in everyday care. Across more than 3.6 million individuals, it detects cancer, infers tumor stage, and stratifies treatment response and recurrence risk, pointing toward risk-adapted cancer care built on data already in hand.

arXiv:2609.12818v1 Announce Type: cross Abstract: Long video understanding often behaves like a visual needle-in-a-haystack problem: query-relevant evidence is sparsely distributed across long temporal spans, while packing dense frames into a single VLM context incurs \textit{context rot} and high cost. Existing video agents often rely on query-agnostic offline preprocessing or ad hoc tool sets, which can miss query-specific details and waste computation. In this work, we present VideoXAgent, a purely online video-agent harness for long video understanding that starts from the given video file and user query, plans and decomposes the task, invokes specialized expert tools on demand, and aggregates multimodal evidence to produce a final answer while resolving conflicts among observations. To support this on-demand invocation, we design a suite of heterogeneous expert tools guided by a data-driven taxonomy of atomic capabilities, spanning scripts, VLMs, and domain models (e.g., detection, OCR, ASR, face recognition). The harness further enforces objective evidence prompting and budget-aware control to curb hallucination and non-termination. Across Video-MME-Long, LongVideoBench-Long, LVBench, and MINERVA, VideoXAgent is competitive with frontier LMMs and video agents under a smaller context footprint---about 50k tokens of agent context per sample, even on hour-long videos. In particular, on complex video-reasoning benchmarks such as MINERVA, it matches this level while using only about 15\% of the context of a 1,024-frame dense-packing baseline. Notably, the harness remains effective with a visually weak or even text-only orchestrator, suggesting that strong long-video understanding can emerge from progressive agentic evidence seeking rather than from packing the full video into a single context. Project page: https://go-agent-x.github.io/video_agent_harness/

Nature Medicine, Published online: 11 September 2026; doi:10.1038/s41591-026-04621-1

The International Consortium of Digital Twins in Healthcare and Medicine was established to advance medical digital twin technology as a new infrastructure for precision health.

Nature Biotechnology, Published online: 02 September 2026; doi:10.1038/s41587-026-03294-y

DNA recombination in rice is optimized by engineering genomic attachment sites.

arXiv:2604.13088v2 Announce Type: replace-cross Abstract: Reinforcement learning for multi-step reasoning with large language models (LLMs) typically relies on sparse terminal rewards, which creates a poorly conditioned credit-assignment problem: the final feedback is propagated uniformly across all intermediate decisions. This leads to high gradient variance, unstable training, and many ineffective updates, ultimately limiting sustained model improvement. We propose a counterfactual-comparison framework for credit assignment. For each input, the framework samples multiple reasoning trajectories and treats their differences as implicit approximations to alternative decisions. This yields an implicit process-level advantage estimator that converts sparse terminal rewards into step-sensitive learning signals. Building on this framework, we introduce Implicit Behavior Policy Optimization (IBPO), which substantially improves training stability and the performance ceiling on mathematical and code-reasoning benchmarks. Our results point to a promising direction for unlocking the reasoning potential of LLMs.

arXiv:2604.17328v2 Announce Type: replace-cross Abstract: This paper investigates the length problem in sequence-level relative reinforcement learning. We observe that, although existing methods partially alleviate length-related phenomena, a more fundamental issue remains insufficiently characterized: the comparison units used during training lack inherent comparability. Building on this observation, we propose a new perspective: the length problem should not be viewed merely as a loss-scaling or normalization bias, but rather as a \emph{comparison unit construction} problem. We further establish a sample-construction-based training framework that, instead of applying post-hoc corrections to unequal-length responses, proactively constructs equal-length, alignable, and comparable training segments during generation. Within this framework, we propose EqLen, a concrete method applicable to group-relative comparison algorithms such as GRPO, GSPO, and RLOO. Through dual-track synchronous generation, prefix inheritance, and segment masking, EqLen efficiently collects effective equal-length training segments and enables stable

arXiv:2605.05226v2 Announce Type: replace-cross Abstract: The central challenge of reinforcement learning for reasoning lies not only in the sparsity of outcome-level supervision, but more fundamentally in how to transform feedback provided only at the end of a sequence into fine-grained learning signals that can guide intermediate reasoning steps. Existing approaches either rely on outcome-level rewards for sequence-level optimization, which makes precise credit assignment difficult, or depend on externally constructed process supervision, which is costly and difficult to scale sustainably. To address this, we propose a new perspective: reinforcement learning for reasoning can be understood as the problem of internalizing outcome supervision into process supervision. From this perspective, we introduce a supervision-internalization method for reinforcement learning for reasoning, enabling the model to automatically extract process-level learning signals through identifying, correcting, and reusing failed reasoning trajectories, thereby achieving finer-grained policy optimization under outcome-only supervision. We further abstract this idea into a new training paradigm, in which the model continually generates and refines its own internal process supervision during reinforcement learning, opening a new path for fine-grained credit assignment in reinforcement learning for reasoning that differs from externally provided process supervision.

arXiv:2605.16302v2 Announce Type: replace-cross Abstract: Reinforcement learning for multi-step reasoning with large language models (LLMs) typically relies on sparse terminal rewards, which creates a poorly conditioned credit-assignment problem: the final feedback is propagated uniformly across all intermediate decisions. This leads to high gradient variance, unstable training, and many ineffective updates, ultimately limiting sustained model improvement. We propose a counterfactual-comparison framework for credit assignment. For each input, the framework samples multiple reasoning trajectories and treats their differences as implicit approximations to alternative decisions. This yields an implicit process-level advantage estimator that converts sparse terminal rewards into step-sensitive learning signals. Building on this framework, we introduce Implicit Behavior Policy Optimization (IBPO), which substantially improves training stability and the performance ceiling on mathematical and code-reasoning benchmarks. Our results point to a promising direction for unlocking the reasoning potential of LLMs.

arXiv:2410.17954v2 Announce Type: replace Abstract: Sparse Mixture-of-Experts (MoE) models can outperform dense large language models at similar computation by activating only a small set of experts per token. However, stacking many expert modules introduces substantial parameter memory, which makes MoE models difficult to deploy in memory-constrained environments such as single-GPU devices. Offloading alleviates this issue by storing inactive experts in CPU memory and loading them on demand, but existing methods remain limited: static caches disregard input-dependent routing, and methods that train separate models to predict expert usage ahead of time are often inaccurate or require significant training cost. We propose ExpertFlow, a lightweight MoE inference system that addresses this routing dependency through three coordinated components: 1) a transformer-based routing path predictor that estimates expert usage across all MoE layers in a single forward pass, 2) a token scheduler that groups tokens with similar predicted routes to improve expert utilization, and 3) a predictive expert cache that loads only the required experts while correcting mispredictions at runtime. Together, these components enable efficient expert loading and execution, reducing GPU memory usage by up to 93.72% and improving inference throughput by up to 10x over strong offloading baselines on a single GPU.

Nature Cancer, Published online: 03 April 2026; doi:10.1038/s43018-026-01141-2

Li et al. present PRET, a few-shot system for pan-cancer detection not requiring model fine-tuning, validated it in multicenter datasets and found that it outperformed existing approaches across tasks and pathologists in lymph node metastasis detection.

An in vivo, noncanonical ferroptosis pathway, critical for tumor supression, is driven by reactive oxygen species-induced phosphatidic acid peroxidation at the ER under the control of the GPX1-OSBPL8 signaling axis.

arXiv:2511.21732v2 Announce Type: replace-cross Abstract: Humor, as both a creative human activity and a social binding mechanism, has long posed a major challenge for AI generation. Although producing humor requires complex cognitive reasoning and social understanding, theories of humor suggest that it follows learnable patterns and structures, making it theoretically possible for generative models to acquire them implicitly. In recent years, multimodal humor has become a prevalent form of online communication, especially among Gen Z, highlighting the need for AI systems capable of integrating visual understanding with humorous language generation. However, existing data-driven approaches lack explicit modeling or theoretical grounding of humor, often producing literal descriptions that fail to capture its underlying cognitive mechanisms, resulting in the generated image descriptions that are fluent but lack genuine humor or cognitive depth. To address this limitation, we propose HUMORCHAIN (HUmor-guided Multi-step Orchestrated Reasoning Chain for Image Captioning), a theory-guided multi-stage reasoning framework. It integrates visual semantic parsing, humor- and psychology-based reasoning, and a fine-tuned discriminator for humor evaluation, forming an interpretable and controllable cognitive reasoning chain. To the best of our knowledge, this is the first work to explicitly embed cognitive structures from humor theories into multimodal humor generation, enabling a structured reasoning process from visual understanding to humor creation. Experiments on Meme-Image-No-Text, Oogiri-GO, and OxfordTVG-HIC datasets show that HUMORCHAIN outperforms state-of-the-art baselines in human humor preference, Elo/BT scores, and semantic diversity, demonstrating that theory-driven structured reasoning enables large language models to generate humor aligned with human perception.

Background: Short-video platforms, characterized by algorithmic curation and passive consumption, have emerged as dominant components of digital life. However, the associations between short-video platform use and health across different groups and usage behaviors remain understudied. Objective: This study investigates associations between short-video platform use and health, examining whether these relationships vary across health status, usage behaviors, and socioeconomic status. Methods: A cross-sectional study was conducted using multistage stratified sampling across eastern, central, and western China from July to September 2024. The inclusion criteria were age 18 years or older, ability to communicate effectively, and no cognitive disorders or mental disturbance. Of 7725 participants enrolled, 46.96% (n=3628) were male, and the average age was 65.49 (SD 8.39) years. The data were collected via face-to-face interviews using a structured questionnaire. Self-rated health and relative health deprivation (Kakwani index) were used to measure health. Quantile regression explored associations between whether using short-video platform and health varies across the health distribution, while linear regression examined associations of years, frequency, daily duration, and purpose diversity of short-video platform use with health. Moderating effect analysis explored the role of socioeconomic status in the relationship between the daily duration of use and health. Results: Coefficients were tested using 2-tailed tests, and statistical significance was defined as a 2-sided value less than .05. Quantile regression revealed heterogeneous associations. Compared to nonusers, short-video platform users had better self-rated health at the 70th to 90th quantiles and lower relative health deprivation at the 10th to 30th quantiles. However, the users at the 10th quantile of self-rated health had worse self-rated health (=−2.224, 95% CI −3.835 to −0.613). Longer engagement (≥3 y) correlated with lower relative health deprivation (=1.970, 95% CI 0.308-3.632), while daily use of 1‐4 hours was associated with poorer self-rated health (=−3.385, 95% CI −4.872 to −1.898; =−3.038, 95% CI −5.054 to −1.022) and higher relative health deprivation (=0.035, 95% CI 0.021-0.050;

arXiv:2603.08572v1 Announce Type: cross Abstract: Learning natural, stable, and compositionally generalizable whole-body control policies for humanoid robots performing simultaneous locomotion and manipulation (loco-manipulation) remains a fundamental challenge in robotics. Existing reinforcement learning approaches typically rely on a single monolithic policy to acquire multiple skills, which often leads to cross-skill gradient interference and motion pattern conflicts in high-degree-of-freedom systems. As a result, generated behaviors frequently exhibit unnatural movements, limited stability, and poor generalization to complex task compositions. To address these limitations, we propose MetaWorld-X, a hierarchical world model framework for humanoid control. Guided by a divide-and-conquer principle, our method decomposes complex control problems into a set of specialized expert policies (Specialized Expert Policies, SEP). Each expert is trained under human motion priors through imitation-constrained reinforcement learning, introducing biomechanically consistent inductive biases that ensure natural and physically plausible motion generation. Building upon this foundation, we further develop an Intelligent Routing Mechanism (IRM) supervised by a Vision-Language Model (VLM), enabling semantic-driven expert composition. The VLM-guided router dynamically integrates expert policies according to high-level task semantics, facilitating compositional generalization and adaptive execution in multi-stage loco-manipulation tasks.

arXiv:2603.03714v1 Announce Type: cross Abstract: We study a systematic bias in modern image generation models: the mention order of entities in text spuriously determines spatial layout and entity--role binding. We term this phenomenon Order-to-Space Bias (OTS) and show that it arises in both text-to-image and image-to-image generation, often overriding grounded cues and causing incorrect layouts or swapped assignments. To quantify OTS, we introduce OTS-Bench, which isolates order effects with paired prompts differing only in entity order and evaluates models along two dimensions: homogenization and correctness. Experiments show that Order-to-Space Bias (OTS) is widespread in modern image generation models, and provide evidence that it is primarily data-driven and manifests during the early stages of layout formation. Motivated by this insight, we show that both targeted fine-tuning and early-stage intervention strategies can substantially reduce OTS, while preserving generation quality.

arXiv:2601.12349v2 Announce Type: replace-cross Abstract: Large multimodal model powered GUI agents are emerging as high-privilege operators on mobile platforms, entrusted with perceiving screen content and injecting inputs. However, their design operates under the implicit assumption of Visual Atomicity: that the UI state remains invariant between observation and action. We demonstrate that this assumption is fundamentally invalid in Android, creating a critical attack surface. We present Action Rebinding, a novel attack that allows a seemingly-benign app with zero dangerous permissions to rebind an agent's execution. By exploiting the inevitable observation-to-action gap inherent in the agent's reasoning pipeline, the attacker triggers foreground transitions to rebind the agent's planned action toward the target app. We weaponize the agent's task-recovery logic and Android's UI state preservation to orchestrate programmable, multi-step attack chains. Furthermore, we introduce an Intent Alignment Strategy (IAS) that manipulates the agent's reasoning process to rationalize UI states, enabling it to bypass verification gates (e.g., confirmation dialogs) that would otherwise be rejected. We evaluate Action Rebinding Attacks on six widely-used Android GUI agents across 15 tasks. Our results demonstrate a 100% success rate for atomic action rebinding and the ability to reliably orchestrate multi-step attack chains. With IAS, the success rate in bypassing verification gates increases (from 0% to up to 100%). Notably, the attacker application requires no sensitive permissions and contains no privileged API calls, achieving a 0% detection rate across malware scanners (e.g., VirusTotal). Our findings reveal a fundamental architectural flaw in current agent-OS integration and provide critical insights for the secure design of future agent systems. To access experimental logs and demonstration videos, please contact yi_qian@smail.nju.edu.cn.