❌

Normal view

Received — 27 May 2026 ⏭

Accelerating Long-Tail Generation in Synchronous RLHF Training via Adaptive Tensor Parallelism

✇cs.AI, q-bio.NC updates on arXiv.org
By: Long Zhao · Qinghe Wang · Jiaan Zhu · Youhui Bai · Zewen Jin · Chaoyi Ruan · Shengnan Wang · Cheng Li
26 May 2026 at 12:00
arXiv:2605.23945v1 Announce Type: new Abstract: Reinforcement Learning from Human Feedback (RLHF) has become a key post-training paradigm for improving model quality. However, the synchronous three-stage RLHF pipeline is often bottlenecked by the generation stage, where response-length skew causes the effective batch size to shrink rapidly during decoding, leaving GPUs underutilized while a few long responses remain unfinished. Mainstream frameworks employ a static tensor parallelism (TP) configuration that cannot adapt to changing batch characteristics, leaving substantial performance headroom unexplored. We propose PAT, an adaptive TP method that dynamically reconfigures TP during the generation stage of each RLHF iteration. PAT introduces two key techniques. First, a predictor-guided online reconfiguration method decides both the reconfiguration point and the target TP configuration based on offline profiling, triggering reconfiguration only when the predicted latency benefit outweighs the reconfiguration overhead. Second, a lightweight online reconfiguration mechanism updates only the states and layouts affected by TP changes: it adapts unfinished decoding states through a cost-model-based choice between KV-cache migration and recomputation, performs in-place weight resharding, and reuses cached communication groups. We implement PAT on top of SGLang and integrate it with the VeRL framework. Evaluations on LLaMA3.1-8B and Qwen3-14B using DeepScaleR show that PAT reduces generation latency by up to 34.6% and end-to-end RLHF training iteration latency by up to 27.2% compared to the original VeRL setup.

IndexMem: Learned KV-Cache Eviction with Latent Memory for Long-Context LLM Inference

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xintong Yang · Hao Gu · Binxing Xu · Lujun Li · Bei Liu · Jiacheng Liu · Qiyuan Zhu · Sirui Han · Yike Guo
26 May 2026 at 12:00
arXiv:2605.25475v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly expected to operate over long contexts, yet standard softmax attention incurs a KV cache that grows linearly with sequence length, quickly becoming the bottleneck for long context inference. A practical remedy is to evict less important KV entries; however, existing eviction policies are largely heuristic and struggle to capture the rich, input-dependent distribution of token importance. In this work, we introduce a learnable indexer that predicts KV importance, enabling more accurate retention of critical tokens. Meanwhile, naively evicting tokens permanently discards their information, leading to irreversible forgetting and degraded retrieval over long ranges. To address this, we propose a lightweight latent memory module that compresses evicted tokens into a compact, online-updated state and provides residual readouts to compensate for the attention contributions lost through KV eviction. Collectively, our method enables accurate long-context inference under a bounded KV budget, delivering consistent improvements on RULER (4K/16K) across Qwen, Mistral, and Llama models (up to 25 points under aggressive eviction), markedly more stable Needle-in-a-Haystack retrieval, and superior LongBench scores and compression curves compared to existing eviction policies.

FG-CLIP 2: A Bilingual Fine-grained Vision-Language Alignment Model

✇cs.AI, q-bio.NC updates on arXiv.org
By: Chunyu Xie · Bin Wang · Fanjing Kong · Jincheng Li · Dawei Liang · Ji Ao · Dawei Leng · Yuhui Yin
26 May 2026 at 12:00
arXiv:2510.10921v3 Announce Type: replace-cross Abstract: Fine-grained vision-language understanding requires precise alignment between visual content and linguistic descriptions, a capability that remains limited in current models, particularly in non-English settings. While models like CLIP perform well on global alignment, they often struggle to capture fine-grained details in object attributes, spatial relations, and linguistic expressions, with limited support for bilingual comprehension. To address these challenges, we introduce FG-CLIP 2, a bilingual vision-language model designed to advance fine-grained alignment for both English and Chinese. Our approach leverages rich fine-grained supervision, including region-text matching and long-caption modeling, alongside multiple discriminative objectives. We further introduce the Textual Intra-modal Contrastive (TIC) loss to better distinguish semantically similar captions. Trained on a carefully curated mixture of large-scale English and Chinese data, including a newly released 12M Chinese region-text dataset, FG-CLIP 2 achieves powerful bilingual performance. To enable rigorous evaluation, we present a new benchmark for Chinese multimodal understanding, featuring long-caption retrieval and bounding box classification. Extensive experiments on 29 datasets across 8 tasks show that FG-CLIP 2 outperforms existing methods, achieving state-of-the-art results in both languages. We release the model, code, and benchmark to facilitate future research on bilingual fine-grained vision-language alignment.

E3AD: An Emotion-Aware Vision-Language-Action Model for Human-Centric End-to-End Autonomous Driving

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yihong Tang · Haicheng Liao · Tong Nie · Junlin He · Ao Qu · Kehua Chen · Wei Ma · Zhenning Li · Lijun Sun · Chengzhong Xu
26 May 2026 at 12:00
arXiv:2512.04733v2 Announce Type: replace-cross Abstract: End-to-end autonomous driving (AD) systems increasingly adopt vision-language-action (VLA) models, yet they typically ignore the passenger's emotional state, which is central to comfort and AD acceptance. We introduce Open-Domain End-to-End (OD-E2E) autonomous driving, where an autonomous vehicle (AV) must interpret free-form natural-language commands, infer the emotion, and plan a physically feasible trajectory. We propose E3AD, an emotion-aware VLA framework that augments semantic understanding with two cognitively inspired components: a continuous Valenc-Arousal-Dominance (VAD) emotion model that captures tone and urgency from language, and a dual-pathway spatial reasoning module that fuses egocentric and allocentric views for human-like spatial cognition. A consistency-oriented training scheme, combining modality pretraining with preference-based alignment, further enforces coherence between emotional intent and driving actions. Across real-world datasets, E3AD improves visual grounding and waypoint planning and achieves state-of-the-art (SOTA) VAD correlation for emotion estimation. These evaluation results show that injecting emotion into VLA-style driving yields more human-aligned grounding, planning, and feedback.

RecGOAT: Graph Optimal Adaptive Transport for LLM-Enhanced Multimodal Recommendation with Dual Semantic Alignment

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yuecheng Li · Hengwei Ju · Zeyu Song · Wei Yang · Chi Lu · Peng Jiang · Kun Gai
26 May 2026 at 12:00
arXiv:2602.00682v2 Announce Type: replace-cross Abstract: Integrating large language model (LLM) representations into multimodal recommendation has shown promise, yet a fundamental challenge remains largely overlooked: the semantic heterogeneity between generative LM representations and the ID-based collaborative signals that recommendation systems rely on. Naively injecting LM features without alignment degrades recommendation performance rather than improving it. To resolve this, we propose RecGOAT, a dual-granularity semantic alignment framework built on graph neural networks and optimal transport theory. RecGOAT first enriches collaborative semantics through multimodal attentive graphs that capture item-item, user-item, and user-user relationships, initializing user representations via LLM-inferred behavioral preferences. It then aligns LM-derived modality representations with recommendation IDs at two complementary granularities: (1) instance-level alignment via cross-modal contrastive learning (CMCL), which produces discriminative per-sample representations; and (2) distribution-level alignment via optimal adaptive transport (OAT), which minimizes the 1-Wasserstein distance between ID distributions and LLM semantics to produce a unified, consistently aligned feature space. Theoretically, we prove that the unified representation achieves strictly lower target error than any single-modality representation, with the gap bounded by the Wasserstein distance and the InfoNCE loss, providing rigorous guarantees for both alignment consistency and fusion comprehensiveness. Extensive experiments on three public benchmarks demonstrate state-of-the-art performance. Deployment on a large-scale online advertising platform further validates RecGOAT's industrial scalability. Our code is available at https://github.com/6lyc/RecGOAT-LLM4Rec.

A pathogen lncRNA secreted into rice sequesters a host miRNA for virulence

✇Nature - Issue - nature.com science feeds
By: Min He · Jia Su · Xiaogang Zhou · Tuo Qi · Jiazheng Wang · Tianxin Zhang · Jinhua Chen · Mawsheng Chern · Youpin Xu · Xiang Lu · Qingqing Hou · Hongrui Liu · Wenfang Huang · Jiawei Liu · Shiying Li · Yunfei Zhu · Xue Chen · Senyu Ran · Han Li · Renju Liu · Mingliang Lei · Guosong Shu · Haicheng Liao · Junjie Yin · Xiaobo Zhu · Yongyan Tang · Li Song · Long Wang · Qing Xiong · Jiali Liu · Yu Bi · Yihua Yang · Hai Qing · Bingtian Ma · Yan Li · Jing Fan · Peng Qin · Yuping Wang · Jianming Wu · Wenming Wang · Shigui Li · Weitao Li · Jing Wang · Xuewei Chen
20 May 2026 at 08:00

Nature, Published online: 20 May 2026; doi:10.1038/s41586-026-10572-x

A fungal long non-coding RNA from Magnaporthe oryzae translocates into rice cells to sequester a host microRNA that normally represses PKR1, a negative immunity regulator, thereby facilitating infection and revealing a widespread RNA-based pathogen–host interaction mechanism.

A SAUR gene enhances maize drought resilience by promoting silk elongation

✇Nature - Issue - nature.com science feeds
By: Chaohui Zhu · Zhirui Yang · Shiping Yang · Xueyan Zhou · Boxin Liu · Tian Tian · Bochen Zhao · Yingying Xie · Yujun Liu · Jinkui Cheng · Huaijun Tang · Yanjun Zhang · Xiaoqing Xie · Lei Zhang · Cheng Liu · Xingrong Wang · Shengxue Liu · Feng Qin
20 May 2026 at 08:00

Nature, Published online: 20 May 2026; doi:10.1038/s41586-026-10566-9

The Small Auxin Up RNA (SAUR) protein ZmSAUR72 in maize (Zea mays) promotes silk growth via regulation of H+-ATPase activity, and is a key determinant of the anthesis-silking interval and thus resilience to drought.

Multi-omics integration identifies ribosome biogenesis-active macrophage subpopulation and its key gene GNL2 in driving liver hepatocellular carcinoma progression and mechanisms

✇Omics in Hepatocellular
By: Yajie Qi · Kun Li · Pincheng Li · Jianyu Yan · Shuyue Feng · Dan Wan · Ke Du · Xiao Liang · Fan Yang · Erzheng Zhou · Na Huang · Qian Wang · Nanbin Liu
14 May 2026 at 18:00

Cancer Cell Int. 2026 May 14. doi: 10.1186/s12935-026-04330-2. Online ahead of print.

ABSTRACT

BACKGROUND: Liver hepatocellular carcinoma (LIHC) is a common malignancy, yet the core genes driving its progression and potential therapeutic targets remain insufficiently explored. Ribosome biogenesis (RB) is a critical biological process linked to various cancers; however, its systematic role in LIHC remains unclear.

METHODS: This study integrated LIHC single-cell RNA-Seq, bulk RNA-Seq, and spatial transcriptomic data with ribosome biogenesis-related gene sets to construct a single-cell atlas of LIHC. Weighted Gene Co-expression Network Analysis (WGCNA) was employed to characterize myeloid cell subsets. Furthermore, an LIHC prognostic risk model based on RB-related genes was developed using 117 machine-learning algorithm combinations. Key findings were subsequently corroborated through experimental validation and clinical sample analysis.

RESULTS: We identified a distinct macrophage subpopulation with high ribosome biogenesis activity, termed ribosome biogenesis-active macrophages (RAMs). These cells exhibited strong communication with inflammatory macrophages, potentially mediated by MIF-related receptor-ligand interactions. We further constructed an 8-gene prognostic model (PA2G4, GNL2, PWP1, DDX49, NOC4L, GDI2, CST7, and RCL1), which showed good predictive performance. Drug sensitivity analysis suggested that the high-risk group may be more responsive to several agents, including docetaxel. Among these genes, GNL2 was selected for further investigation. Elevated GNL2 expression was associated with increased stemness features in myeloid cells. Molecular docking analysis identified several candidate compounds with potential binding affinity to GNL2. Functionally, GNL2 knockdown in macrophages reduced TGF-β and TNF-α expression and was associated with decreased proliferation, migration, and invasion of LIHC cells.

CONCLUSION: We identified a highly active ribosome biogenesis-macrophage subpopulation (RAM), and constructed a robust risk model to aid in the diagnosis, prognosis, and treatment of LIHC. GNL2 is associated with increased expression of TGF-β and TNF-α and may contribute to LIHC progression.

PMID:42135716 | DOI:10.1186/s12935-026-04330-2

There are no more articles
❌