Jina-OCR-v1 is an end‑to‑end document parsing model designed for low‑budget GPUs, combining a compressed‑vision encoder with a 3B mixture‑of‑experts decoder that activates about 570 M parameters per token. It uses a FastMTP speculative decoding head that shares a single draft block across three prediction steps, with greedy verification ensuring lossless decoding. Post‑training includes instruction alignment, robustness fine‑tuning on difficult documents, and GRPO with dense verifiable rewards, achieving 91.14 on OmniDocBench v1.6 and 83.4 on olmOCR‑Bench while delivering the highest page throughput at 2.57 pages per second on an NVIDIA L4 GPU.
By Alejandro Bar\'on Garc\'ia, Feng Wang, Emilia Garcia Casademont, Han Xiao
M3T introduces a discrete multi‑modal motion token system for sign language production, addressing the need for non‑manual features such as mouthings, eyebrow raises, gaze, and head movements. The approach couples FLAME’s expressive facial space with SMPL‑X body parameters and uses modality‑specific Finite Scalar Quantization VAEs to achieve high face codebook utilization (99.0%). Trained with an autoregressive transformer and a sign‑to‑text translation objective, M3T outperforms existing methods on three standard datasets, notably improving accuracy on NMFs‑CSL from 49.0% to 58.3% without large‑scale pre‑training.
By Alexandre Symeonidis-Herzig, Jianhe Low, Ozge Mercanoglu Sincan, Richard Bowden
The paper introduces a physics‑informed graph attention network that directly operates on the tetrahedral mesh used in TCAD simulations of FinFET devices. By predicting electrostatic potential and quasi‑Fermi levels at every mesh node and training with both data loss and finite‑volume current‑continuity residuals, the surrogate retains the underlying carrier‑transport physics while achieving size generalization. Benchmarks against Sentaurus Device show sub‑volt RMSE for the drift‑diffusion fields and a per‑design throughput that is orders of magnitude faster, enabling rapid Pareto‑front exploration of large multi‑fin arrays that would otherwise be prohibitively slow to simulate.
By Leonid Popryho, Ayoub Sadeghi, Inna Partin-Vaisband
GrowPage is an on‑demand key–value (KV) budgeting framework designed to improve large language model (LLM) reasoning serving. It treats KV capacity as a runtime resource, using lightweight dual‑timescale query summaries to track recent and long‑term attention patterns and estimate demand evolution. At each capacity boundary, GrowPage either compresses KV states within the current allocation or acquires an additional physical page, integrating with PagedAttention’s page‑level memory abstraction to maintain continuous batching and prefix caching.
By Qiankun Ma, Yanjiang Zhou, Zinan Xiong, Haofei Wang, Zhen Song, Yang Xiang, Ziyao Zhang, Hairong Zheng
The paper investigates how model compression impacts test-time adaptation (TTA) in deep neural networks. Using ResNet‑18 and ViT‑Base on CIFAR‑10‑C and ImageNet‑C, the authors evaluate several compression techniques alongside standard TTA methods, introducing a diagnostic framework to assess representational expressivity and adaptation subspace compatibility. Results show that while compressed models maintain high accuracy under supervised adaptation, their TTA performance deteriorates with increased compression due to reduced representational diversity and structural constraints that limit recoverability.
By Francesco Corti, Dong Wang, Young D. Kwon, Cecilia Mascolo, Olga Saukh
VestigeKV is a new KV‑cache technique that uses a 64‑dimensional vestigial branch—originally a RoPE component repurposed during NoPE training—as a query‑independent eviction signal. By reading only 11 % of each cache row, the method partitions the cache into an attended tier (top‑m rows) and an archive tier (all other rows), which is GPU‑resident and never deleted. The approach achieves near‑perfect retrieval (1.00 at 8×, 0.92 at 32×) without any training, quantization, or changes to weights or kernels.
By WenJie Fan
The paper presents a consensus framework for Bayesian Network fusion that balances dependency preservation with computational tractability by enforcing treewidth constraints. It introduces genetic algorithms featuring advanced initialization, specialized operators, and a tailored fitness function to prioritize shared structures among input networks. Experiments on synthetic and real-world BNs demonstrate that these genetic algorithms outperform adapted methods and greedy baselines.
By Pablo Torrijos, Jos\'e A. G\'amez, Jos\'e M. Puerta, Juan A. Aledo
The paper introduces the concept of a free pause token, which allows a language model to perform additional computation for each next-token prediction without increasing the sequence length or KV cache usage. This pause token operates as a parallel prediction stream over a shared backbone, improving next-token accuracy by 2–3 centi‑nats on a 1‑billion‑parameter model. While it adds minimal inference overhead, the primary cost lies in training, where compute requirements rise modestly (up to 1.14×) yet most benefits are retained.
By John Langford, Nathan Godey, Giovanni Monea, Yoav Artzi, Harry Dong, Ying Fan, Gustavo de Rosa, Zheng Zhan
FlashRender is a few-step generative rendering framework that quickly retakes a source video along a target camera trajectory. It addresses discretization error by introducing Representation Transformation and Alignment (RETA) to align source-video representations with target-video features, reducing denoising trajectory curvature. The model is further refined with a MeanFlow objective and on-policy flow map distillation, achieving video quality and geometric consistency comparable to multi-step baselines at a 25× lower sampling cost while improving camera controllability.
By Byeongjun Park, Byung-Hoon Kim, Hyungjin Chung
SGD-KV is a head‑aware framework for compressing key‑value caches in large language models. It uses a chunk‑summarization diagnostic task to identify attention heads that specialize in hierarchical information aggregation, allowing the KV cache budget to be allocated based on each head’s summarization score. Experiments on Qwen2.5‑7B‑1M and Qwen3‑32B show state‑of‑the‑art performance on up to 1M‑token contexts while cutting KV cache memory usage by up to 75%.
By Zeyu Liu, Woomin Song, Xuandi Fu, Sai Muralidhar Jayanthi, Vivek Govindan, Aram Galstyan, Sravan Babu Bodapati, Srikanth Ronanki
The paper introduces AdaptiveSpec, a training‑free speculative decoding method that simultaneously adapts the per‑step verification rule and the draft‑tree shape using signals generated during decoding. It replaces the fixed token‑match rule with a margin‑based threshold and adjusts tree depth, width, and node count based on draft confidence and recent acceptance history, allowing the total draft count to vary. Experiments on SGLang show up to 56% throughput gains over EAGLE‑3 while maintaining 93% of lossless task accuracy on GSM8K, MATH‑500, and HumanEval across three models.
By Oszk\'ar Urb\'an, Young D. Kwon, Stylianos I. Venieris, Cecilia Mascolo
CoMAP introduces a framework that jointly evolves textual world models and agent policies through a closed‑loop interaction. At each decision step the world model forecasts future state feedback for candidate actions, while the agent reflects on the reliability of this feedback to refine its action. The resulting on‑policy trajectories are used to self‑distill and update the world model, improving prediction accuracy and long‑horizon decision‑making across embodied planning, web navigation, and tool‑use benchmarks.
By Youwei Liu, Jian Wang, Hanlin Wang, Wenjie Li
ProgResViT is an input‑adaptive Vision Transformer that processes images progressively across multiple rounds, starting with a low‑resolution image and a narrow subnetwork and refining the prediction with higher resolution and a wider subnetwork if needed. The method introduces Progress‑Conditioned Soft Gating (PSG) to share a single backbone across rounds while conditioning token fusion and layer outputs on the current round, block, and input resolution. Experiments on DeiT show improved accuracy‑compute trade‑offs compared to adaptive‑width, adaptive‑depth, and dynamic‑token baselines, and the design also benefits self‑supervised DINO representations and downstream semantic segmentation.
By Ali Hojjat, Janek Haberer, Olaf Landsiedel
Tree-VQ introduces a progressive tree‑structured vector quantization framework for learned image compression, organizing discrete codewords in a hierarchical binary tree where each latent token is represented by a routed root‑to‑leaf path. Every prefix of this path yields a valid quantized representation, enabling coarse reconstructions from shallow nodes and successive refinements from deeper nodes. The method incorporates a prefix‑compatible tree entropy model, rate‑aware refinement scheduling, and hierarchical prefix supervision to achieve efficient, low‑latency compression with superior perceptual quality and fewer parameters compared to existing approaches.
By Xinkun Wang, Tianyi Xu, Qingyu Luo, Mingming Ma, Changzhe Jiao, Fu Li, Yi Niu
The paper introduces LRE (Learned Relevance Eviction), a lightweight, CPU‑only, language‑model‑free scorer that learns which parts of an agent’s interaction history are task‑critical and preserves them verbatim. In experiments, LRE matches or surpasses baseline eviction policies on accuracy‑cost trade‑offs, recovers 93% of full‑history accuracy, reduces worst‑case prompt size by 52%, and outperforms dense and token‑pruning encoders in conversational memory while being 295–1569× smaller. The method also achieves superior budgeted answer quality on LoCoMo reading and can be trained annotation‑free, recovering 95% of supervised scorer performance.
By Nusrat Jahan Lia, Aritra Mazumder
The paper investigates on‑policy distillation (OPD) of large language models when trained on a minimal amount of data, specifically a single query. It finds that one‑shot OPD continues to improve for hundreds of steps and recovers most of the performance gains achieved with full‑dataset training across various tasks and model families. The authors attribute this to the high state coverage achieved early in training and show that adding a few semantically distinct queries further boosts coverage and accuracy, eventually matching full‑data results. They conclude that OPD is data‑overfed but algorithm‑starved, highlighting the need to focus on step efficiency and the underlying mechanisms of its recent successes.
By Zixuan Fu, Bingxiang He, Yuxin Zuo, Haohuan Huang, Jinqian Zhang, Ruhang Xiao, Cheng Qian, Qinyu Luo, Huan-ang Gao, Yudong Wang, Zhiyuan Liu, Ning Ding, Chaojun Xiao
The paper introduces Modality‑Specific Frequency Distillation (MSFD), a continual learning framework for video deepfake detection that separates spatial, temporal, and spatiotemporal features in the frequency domain. By preserving each modality independently and applying a cross‑modality decorrelation loss, MSFD adapts to new forgery patterns while maintaining performance across diverse continual deepfake video scenarios. Experiments demonstrate that this approach outperforms state‑of‑the‑art methods in both adaptation and retention.
By Taehoon Kim, Jongwook Choi, Heejae Jo, Byungmin Park, Jongwon Choi
EmoDistill is an offline framework that distills emotional negotiation skills from large language model interactions into smaller agents. It separates emotion selection, handled by an Implicit Q‑Learning selector, from emotion‑conditioned expression, learned by a LoRA‑adapted 7B policy via supervised fine‑tuning and judge policy optimization. Experiments across four negotiation domains show that the full EmoDistill policy outperforms vanilla and IQL‑only baselines, while removing the explicit emotion channel markedly reduces negotiation utility and reveals partial, domain‑dependent transfer to unseen counterparties.
By Yunbo Long, Haolang Zhao, Lukas Beckenbauer, Liming Xu, Alexandra Brintrup
The paper presents a 4‑bit quantization recipe, Minima: NVFP4 W4A4, that fully quantizes all linear layers—including the Gated DeltaNet (GDN) recurrent blocks—of the 27‑billion‑parameter Qwen3.8 LLM. Across a suite of benchmarks (perplexity, MMLU‑Pro, GSM8K, AIME'25, GPQA‑Diamond, LiveCodeBench, and RULER retrieval), the quantized model matches BF16 performance within seed noise while being 17.5 GiB in size and 14–19 % faster at prefill. The authors attribute this success to four mechanisms: block‑scaling of residuals, robust gate projections, the delta‑rule recurrence’s noise‑plateau behavior, and the per‑token quantization cost’s dilution over long contexts.
By Sergii Kozyrev, Davyd Maiboroda
The paper investigates how on-policy distillation (OPD) and reinforcement learning with verifiable rewards (RLVR) can be combined for post‑training reasoning in large language models. It shows that a two‑stage approach—first applying OPD, then RL—outperforms single‑signal methods and other joint baselines on logic and math reasoning benchmarks. The authors explain this advantage through pass@k analysis, learning dynamics, and parameter updates, concluding that OPD expands solution coverage while RL sharpens performance within that support, and that the OPD validation score is the key trigger for switching to RL.
By Boyan Li, Bingsen Chen, Chenghao Yang, Ping Nie, Chen Zhao, Xi Ye