Vagdhenu is a Sanskrit shloka‑to‑chant text‑to‑speech system that preserves meter (vrutta) and phonological nuances. It builds on an off‑the‑shelf flow‑matching backbone and a large‑scale neural vocoder, adding a Kannada‑based frontend to avoid schwa deletion, a phonology‑aware frontend handling visarga sandhi and sibilant distinctions, and a vrutta‑aware reference selection mechanism. The authors report that a text‑side prosody conditioner is ineffective in their architecture, while reference clips and voice‑steering retraining provide the necessary prosody control, and they demonstrate the system’s performance on a 32‑chapter video corpus and an audio app covering 18,000 verses.
whyItMatters":"The system delivers high‑fidelity, meter‑aware Sanskrit chanting, enabling large‑scale deployment of authentic recitations for educational and cultural preservation purposes."
By Prathosh A P
ET‑Prune is a training‑free framework that dynamically allocates visual token budgets in multimodal large language models based on question‑conditioned evidence. It protects text‑like spatial regions, converts evidence uncertainty into a token floor, and progressively prunes concentrated evidence while retaining diffuse or text‑dense tokens. In six backbone‑benchmark comparisons, ET‑Prune matches or outperforms other pruned methods while keeping roughly half the visual tokens, achieving notable gains on OCRBench‑v2 and MMBench v1.1.
By Zizhong Ding, Junxian Li, Kai Liu, Shaoqiu Zhang, Xiao Xiao, Linghe Kong, Yulun Zhang
ADeptS-Bench is a new benchmark designed to assess the trustworthiness of Computer Use Agents (CUAs) across mobile and desktop devices. It consists of two streams: a Safety stream with paired benign and malicious tasks that embed visual threats, and a Disambiguation stream that tests whether agents seek clarification when instructions are ambiguous. Evaluation of seven models shows none consistently achieves high task success while keeping attack success low, and all models exhibit problematic behaviors such as unhesitant checkout on a $25K order and failure to detect a mislabeled factory reset button.
By Joy Chen, Alejandro Castillejo Munoz, Pierluca D'Oro, Yuxuan Sun, Chloe Evans, Joseph Tighe
Anian is a safety‑gated multimodal AI backend designed for perinatal mental‑health support and mindfulness‑intervention routing. It maps user input into a four‑layer hierarchical state representation—emotion, psychosocial constructs, safety risk, and intervention routes—then fuses local and external risk signals to decide whether to generate AI responses or provide fixed safety content. Prototype evaluation on large public corpora showed high classification performance and perfect high‑risk recall in a controlled stress test, though clinical validity remains unestablished.
By Lei Wang, Xiao Wang, Lei Li
The paper introduces Visual Retrieval Heads (VRHs), a small fraction of attention heads in vision‑language models that are causally responsible for grounding text descriptions to image regions. By recasting head‑scoring methods and evaluating across eleven VLMs and five benchmarks, the authors show that masking the top 20 VRHs can drop grounding accuracy by up to 80 percentage points, while random masking has little effect. VRHs generalize across various visual reference tasks, preserve output format while corrupting localization, and transfer causally across models sharing an LLM backbone.
By Chanho Park, Daehyeon Choi, Jihyun Lee, Minhyuk Sung
SpeechGym is an audio‑native environment that lets two omni‑modal models converse entirely in native audio, eliminating external ASR/TTS and API boundaries while preserving the tasks, tools, and success checks of a standard text‑based agent benchmark. By training end‑to‑end, the framework addresses perceptual failures—such as misheard arguments that cascade into failed calls—and behavioural failures, both of which are automatically labeled for free. Using per‑turn process rewards to overcome reward sparsity, agents trained in SpeechGym transfer to an independent voice benchmark, doubling task success and improving efficiency in turns and tokens.
By Jiajun Fan, Jingyuan Li, Prashanth Gurunath Shivakumar, Jia-Hong Huang, Qi Luo, M. Maruf, Ivan Bulyko, Ge Liu, Roger Ren
ReViCo (Real Visual Correction) is a new benchmark that tests Vision Language Models (VLMs) on the task of correcting text errors in real‑world images, requiring deep understanding of visual text and its context. The study evaluates VLMs using both prompt‑based and targeted training approaches, revealing a significant performance gap between current models and humans. The results show that most VLMs struggle to accurately perceive visual text, leading to frequent correction mistakes, thereby underscoring the need for more robust, text‑aware VLMs.
By Bojun Zhang, Junhong Liang, Feifei Zhai, Fengxian Ji, Yu Zhou
AdaThinking-E introduces a reinforcement learning framework that uses one-token entropy regulation to enable large language models to decide adaptively whether to engage in deep reasoning. By measuring entropy in the predicted probability distribution at key decision tokens, the model learns to explore different thinking strategies during training and converge to confident, efficient decision policies. Experiments show the method improves accuracy on complex tasks while reducing computational overhead on simpler ones across various document reasoning benchmarks.
By Zining Wang, Tongkun Guan, Boming Chen, Zhentao Guo, Jianqiang Liu, Chao Jin, Chen Duan, Kai Zhou, Pengfei Yan, Wei Shen, Xiaokang Yang
MedFG-VQA is a lightweight medical visual question answering framework that uses a memory bank to enhance low‑frequency DCT features and graph‑enhanced cross‑attention for visual‑textual alignment. It introduces Frequency‑Memory Fusion to retrieve and fuse low‑frequency information from a learnable memory bank, and Graph‑Aware Cross‑Attention to refine cross‑modal features via graph convolution. The authors also create SynMed‑VQA, a synthetic dataset of over 2 million QA pairs across nine imaging modalities, and show that MedFG‑VQA matches or outperforms larger models on several biomedical VQA benchmarks while keeping computational costs low.
By Haowen Gu, Gensheng Pei, Zeren Sun, Mingwu Ren, Xiangbo Shu, Yazhou Yao, Fumin Shen
NeoTriFuse is a reliability‑aware multimodal fusion framework designed to predict neonatal mortality risk from bedside monitoring data that suffers from extreme class imbalance, heterogeneous risk factors, multi‑scale temporal dynamics, and significant missingness. The method treats missing data as an explicit reliability signal, dynamically adjusting modality contributions during fusion through gating mechanisms that incorporate static perinatal variables, local‑global temporal encoders, and patient‑level statistical summaries. NeoTriFuse achieves competitive performance (F1 ≈ 0.674, AUROC ≈ 0.945) and ablation studies show that its temporal architecture and patient‑level summary branch are key contributors, with reliability‑aware gating further improving threshold‑dependent metrics under heterogeneous observation completeness.
By Jiyuan Tian, Qincheng Shen, Ye Lin, Yu Gao, Haohui Lu
4DSynth is a controllable procedural system that transforms natural-language descriptions, blueprint masks, or single photographs into editable 4D environments featuring explicit geometry, animated actors, collision-free trajectories, and physics-ready simulation states. The system unifies animation, camera planning, rendering, and task generation within a single geometry-grounded representation, enabling scalable creation of dynamic embodied simulation scenes. Using 4DSynth, the authors built 4DSynth-Nav, an interactive navigation benchmark that demonstrates the reproducibility and tunability of procedural failures across vision‑language models.
By Zehao Qi, Haochen Luo, Jia-Wang Bian, Zeyu Ma, Shuyang Sun
The paper shows that a plain Transformer can serve as an effective graph learner by adding three lightweight modifications: simplified L₂ attention, adaptive RMS normalization, and an MLP-based positional encoding stem. These changes preserve token magnitude and enable the model to achieve high expressivity on graph benchmarks, outperforming more complex graph transformer variants. The results suggest that plain Transformers can act as a unified backbone for multimodal learning across language, vision, and graph domains.
By Liheng Ma, Soumyasundar Pal, Yingxue Zhang, Philip H. S. Torr, Mark Coates
The paper introduces Composed Gait Retrieval (CoGR), a task that retrieves a target gait sequence using a reference sequence and a natural language modification query. To support this, the authors create the first gait-language datasets—Language‑Augmented CCPG and Language‑Augmented CASIA‑B—via an automated annotation pipeline powered by large vision‑language models. They propose ComposeGait, an identity‑anchored composition framework with a Part‑aware Identity Adapter that injects identity tokens into a shared Q‑Former, achieving state‑of‑the‑art retrieval performance on both benchmarks.
By Jingchen Fei, Zengbin Wang, Yukun Liu, Muyi Sun, Shibiao Xu, Man Zhang
PACE introduces a training‑free Condense‑and‑Extract framework that speeds up Vision‑Language Model inference by first adaptively downsampling visual inputs before encoding and then selectively retaining essential tokens during decoding. The Adaptive Pixel Compressor (APC) reduces encoder workload while preserving global context, and the Dynamic Dual‑Attention Extractor (DDAE) keeps task‑critical details by fusing visual and language signals. Applied to Qwen2.5‑VL‑7B, PACE maintains 93.8% of performance using only 10% of visual tokens, achieving a 3.1× speedup in time to first token.
By Junjie Liu, Shengyuan Ye, Xu Chen
NeuronFuzz is a white‑box fuzzing framework that uses internal safety neurons of large language models as continuous feedback for safety evaluation, eliminating the need to generate full model responses during testing. It constructs a SafetyOracle that converts neuron activations into a differentiable safety alarm score, enabling gradient‑based identification of sensitive template positions and fluent, context‑compatible prompt mutations. Evaluated on 21 text and multimodal models, NeuronFuzz achieves a 76‑100% jailbreak discovery rate on five white‑box source models and demonstrates strong zero‑shot transfer to open‑weight and proprietary targets.
By Zhiyuan Xu, Muhammad Firhard Roslan, Joseph Gardiner, Sana Belguith, Lichao Wu
The paper introduces Planning Diffusion Policy Optimization (PDPO), an offline‑to‑online reinforcement‑learning framework that employs a diffusion policy to produce short‑horizon action chunks for robot crowd navigation. PDPO is pretrained on collision‑avoidance demonstrations and fine‑tuned online with PPO, generating five‑step action sequences applied in a receding‑horizon manner. The authors also identify a benchmark artifact where agents can leave the valid domain without explicit boundary constraints, and they mitigate this by treating boundary violations as collisions, leading to improved success rates over strong baselines.
By Wendong Li, Jochen Garcke
MM‑Spectrum is a sparse Mixture‑of‑Experts framework designed to infer molecular structures from multimodal spectroscopic data. It introduces a modality‑aware routing mechanism that exposes spectral identity to the router, along with shared and interaction experts of heterogeneous capacities to capture both modality‑unique and cross‑modal synergistic information while reducing noise interference. Experiments across full‑modality, bimodal, and missing‑modality scenarios show consistent and substantial performance gains, supported by ablation studies and interpretability analyses.
By Hai-tao Yu, Nan Min, Zheng Fang, Hongyu Zhan, Yusen Tan, Yuhan Wang, Jun Xia
The paper introduces a method to improve audio‑visual speech recognition by applying contrastive decoding (CD) that contrasts audio‑only with audio‑visual conditioning within the same model. It addresses the issue of a fixed CD strength by scaling the influence adaptively for each token, using reliability signals from attention dynamics and predictive divergence. Experiments on the LRS3 dataset demonstrate consistent gains in both clean and low‑SNR scenarios.
By YoungChae Kim, Da-Hee Yang, Joon-Hyuk Chang
The paper introduces MDiTFace, a diffusion transformer designed for high‑fidelity mask‑text collaborative facial generation. It unifies tokenization of semantic masks and text, enabling synchronous multimodal feature interaction via stacked multivariate transformer blocks. A novel decoupled attention mechanism separates dynamic and static computations, allowing caching of static features and reducing mask‑condition overhead by over 94% while preserving performance, leading to superior facial fidelity and conditional consistency compared to existing methods.
By Yushe Cao, Dianxi Shi, Xing Fu, Xuechao Zou, Haikuo Peng, Xueqi Li, Chun Yu, Junliang Xing
The paper introduces Temporally-Grounded Language Generation (TGLG), a benchmark that tests vision‑language models on their ability to produce semantically accurate and temporally precise utterances in real‑time settings. It identifies perceptual updating and contingency awareness as key capabilities, curates datasets from sports broadcasting and egocentric interactions, and proposes the TRACE metric to jointly evaluate semantic similarity and temporal alignment. The authors also present VLM‑TSI, a model that interleaves visual and linguistic tokens in a time‑synchronized manner, achieving better performance than a strong baseline yet still showing modest overall results, underscoring the challenge of real‑time VLMs.
By Keunwoo Peter Yu, Joyce Chai