Searcharxiv⌕ Search

arXiv subjects

Yiming Wang

Publications and source records attributed to Yiming Wang.

At least 19 recordsLinked to original sources
arXiv2026

VT-MUSE: Multimodal Unified Sequential Visuotactile Representation Learning for Manipulation

We propose VT-MUSE, a Multimodal Unified SEquential representation learning framework for visuotactilemanipulation. Existing approaches often encode visual and tactile observations independently before fusion, limiting their ability to capture fine-grained cross-modal dependencies. Moreover, most methods focus on observations at the current time step and overlook the temporal evolution of contact. VT-MUSE addresses both limitations through a two-stage representation learning framework. In Stage I, modality specific encoders are jointly adapted via cross-modal temporal alignment and masked-view consistency. In Stage II, a conditional variational latent model processes masked visual sequences together with full tactile histories. Auxiliary decoders reconstruct the masked recent visual observations and predict tactile depth changes, encouraging the latent representation to retain both global visual context and local contact dynamics. The learned representation is subsequently integrated into a lightweight Transformer policy through gated cross-attention. On the simulation benchmark, VT-MUSE outperforms the strongest baseline evaluated on all tasks by 11 percentage points and also achieves substantial improvements in real-world experiments.

cs.RO↗
arXiv2026

Mitigating Saliency Collapse: Robust Saliency-Aware Long-Text Image-Text Alignment

Long-text image-text alignment is essential for fine-grained scene understanding, yet existing methods mainly focus on extending context length or improving local alignment. We identify an overlooked failure mode, saliency collapse, where models over-rely on text that features visual prominence while underutilizing non-salient context. Consequently, retrieval performance degrades sharply when salient content is reduced, despite that the non-salient context remains informative. To address this, we propose RoSA, a saliency-aware finetuning framework that decomposes image-text pairs into complementary salient and contextual views, enabling joint learning of prominent and contextual semantics. We propose a modality-independent decomposition strategy to combine object-level visual saliency estimation with LLM-based text decomposition to bridge vision and language. We employ auxiliary objectives to explicitly align these views, ensuring robust representation with both salient and contextual features. RoSA greatly improves long-text retrieval and demonstrates strong robustness to saliency collapse across benchmarks. Moreover, our auxiliary objectives are plug-and-play, bringing consistent gains to existing tuning methods. Code and checkpoints are available at: https://github.com/intelligolabs/RoSA.

cs.AI↗
arXiv2026

Compositional Humanoid Manipulation through Isomorphic Teleoperation and Independent Skill Routing

Humanoid manipulation requires reusable arm skills, yet demonstrating every possible skill combination is impractical. We present Humanoid Semantic Mixture-of-Experts Diffusion Policy (HS-MoDP), which composes separately demonstrated arm skills through independent semantic routing over a shared expert pool. This design preserves arm-specific skill control while sharing action-generation parameters across arms. A semantic scheduler integrates the learned policy with an existing locomotion controller through alternating manipulation and repositioning stages, extending the effective workspace and enabling skill reuse across location changes. To acquire demonstrations of whole-body contact manipulation, we develop G1Exo, an isomorphic exoskeleton that provides direct joint-level control of arm configurations. Experiments on a Unitree G1 show stronger skill composition than the baselines. G1Exo enables more reliable contact teleoperation than end-effector VR-IK, and its demonstrations improve contact-skill learning across four policy architectures. For HS-MoDP, switching from VR-IK to G1Exo demonstrations raises full cabinet workflow success from 35% to 65%. All project hardware designs and source code will be released as open source after the review process.

cs.RO↗
arXiv2026

PrismWAM: Decoupling State Representations in World Action Models

World Action Models (WAMs) typically represent the world through latent embeddings from pretrained video models, which entangle heterogeneous aspects of the world, from how it looks (e.g., texture and color) to how it is physically configured (e.g., occupancy and pose). This raises an underexplored question: $\textbf{what should constitute the state of a WAM?}$ We argue that the world state should separate how the world appears from how the world evolves under action: the former captures semantic context, while the latter is governed by physical state. To this end, we propose $\textbf{PrismWAM}$, which decouples physics from semantics while jointly modeling state and action. PrismWAM factorizes an embodied task into a $\mathrm{Semantic\ Planning}\rightarrow\mathrm{Physical\ Rollout}\rightarrow\mathrm{Executable\ Action}$ pipeline: given only the initial RGB observation, semantic planning learns to predict the final goal image, while physical rollout predicts the fine-grained evolution toward this goal in depth space; executable actions are then generated from the depth rollout. Without any embodied action pretraining or additional data augmentation, PrismWAM achieves strong manipulation performance in both simulation and the real world. Notably, PrismWAM even surpasses a broad range of models with large-scale embodied action pretraining (e.g., OpenWAM-$α$). More broadly, we reframe World Action Models from modeling how the world looks to how it responds to action, paving the way for action-centric world models grounded in occupancy, flow, tactile sensing, proprioception, and beyond.

cs.CV↗
arXiv2026

Learning Navigation from Multimodal Demonstrations with Geometric Candidate Modes

Offline navigation demonstrations are multimodal, because one observation often allows several valid actions, such as passing an obstacle on either side. Behavior cloning averages these actions, potentially leading to unsafe routes. Multimodal policies must discover the alternatives from the data, but similar states can lead to different routes whose current actions overlap. In navigation, the alternatives have a geometric cause and can be computed from one observation. We propose a hierarchical neural-network policy that computes candidate modes from local geometry and learns from demonstrations which mode to follow and how. A high-level policy selects a mode using the goal and its previous choice among a set of candidate modes, and a low-level policy turns it into an action. In simulation, the policy succeeds in over 95\% of seen and unseen test cases, compared with at most 43\% and 20\% for four offline learning baselines and 79\% and 67\% for a heuristic that uses the same modes without learning. The simulation-trained policy also transfers to a LimX Tron 1 bipedal robot.

cs.RO↗
arXiv2026

AEIOU: A Unified Defense Framework against NSFW Prompts in Text-to-Image Models

As text-to-image (T2I) models advance and gain widespread adoption, their associated safety concerns are becoming increasingly critical. Malicious users exploit these models to generate Not-Safe-for-Work (NSFW) images using harmful or adversarial prompts, underscoring the need for effective safeguards to ensure the integrity and compliance of model outputs. However, existing detection methods often exhibit low accuracy and inefficiency. In this paper, we propose AEIOU, a defense framework that is adaptable, efficient, interpretable, optimizable, and unified against NSFW prompts in T2I models. AEIOU extracts NSFW features from the hidden states of the model's text encoder, utilizing the separable nature of these features to detect NSFW prompts. The detection process is efficient, requiring minimal inference time. AEIOU also offers real-time interpretation of results and supports optimization through data augmentation techniques. The framework is versatile, accommodating various T2I architectures. Our extensive experiments show that AEIOU significantly outperforms both commercial and open-source moderation tools, achieving over 95\% accuracy across all datasets and improving efficiency by at least tenfold. It effectively counters adaptive attacks and excels in few-shot and multi-label scenarios.

cs.CR↗
arXiv2026

Training a Compact, Strong, and Realistic Text-to-Speech Model

Text-to-speech (TTS) systems are conventionally optimized for intelligibility and speaker similarity. However, neither objective captures the prosodic detail that distinguishes human performances from synthetic speech, leaving a realism gap that larger models and higher-quality data can narrow but not close. In the absence of an objective that directly targets this gap, realism remains a persistent challenge in TTS. We present Audio8-TTS, a compact 0.6B-parameter model for zero-shot voice cloning that achieves state-of-the-art performance in intelligibility and speaker similarity, with particularly strong results in perceptual naturalness. To align the training signal with human perception rather than signal-level fidelity, we construct a quality ladder from widely used open-source TTS systems and train a realism critic to assess generated speech against matched human recordings of the same sentences. An identity-erasure objective discourages the critic from relying on generator fingerprints. We train the model in three stages: multilingual supervised fine-tuning; group relative policy optimization (GRPO) with word error rate (WER) and speaker similarity (SIM) rewards; and GRPO guided by the realism critic. The final stage improves perceived prosodic naturalness and speaker similarity while preserving WER performance.

cs.SD↗
arXiv2026

ROAR: Retrieval Opportunity-Aware Refinement for Zero-Shot Time Series Forecasting

Retrieval augmentation provides time series forecasters with historical continuations, yet even candidates that outperform the base forecast may fail to improve the final prediction. We propose ROAR, a Retrieval Opportunity-Aware Refinement framework for zero-shot time series forecasting. To better exploit these improvement opportunities, its training objective allocates additional emphasis across queries based on base-forecast difficulty and the relative improvement offered by retrieved candidates. Using this objective, ROAR first learns to aggregate aligned historical candidates and uses a learned gate to control their correction strength against a fixed base forecaster. It then jointly calibrates the forecasting module and gate to coordinate their contributions, while anchoring the combined prediction to the first-stage refined output. We derive exact decompositions of refinement gains and the opportunity-weighted training loss, and characterize additional gains from joint calibration under a local linearization. Experiments on seven benchmarks show that ROAR achieves the lowest average MSE among the compared methods. Further evaluations demonstrate average forecasting improvements across multiple backbone families and retrieval-augmented forecasters.

cs.LG↗
arXiv2026

PEARS: Physical-Prior-Guided Efficient Adaptation via Failure Reasoning and Diffusion Steering for Tactile Manipulation

Pretrained robotic policies can suffer substantial performance degradation under out-of-distribution (OOD) conditions encountered during deployment, motivating post-training through real-world interaction. However, reinforcement-learning (RL)-based post-training typically requires substantial environment interactions, a burden that is especially significant in manipulation, where each trial can be slow, costly, or destructive. Therefore, we present PEARS, a physics-prior-guided hybrid RL framework for sample-efficient online adaptation of pretrained policies with tactile feedback. After each episode, its physics-guided force reasoning (PFR) module uses physical priors encoded in a vision-language model (VLM) to diagnose failures from the visual outcome and tactile interaction history and update task-appropriate contact-force bounds. A high-frequency hybrid force-position controller then enforces these bounds during contact. Complementarily, tactile-conditioned diffusion steering reinforcement learning adjusts the latent noise of the frozen flow-matching policy to correct errors in free-space motion and contact timing without updating the base model. In simulation, PEARS improves success rates by 12.4-37.4 percentage points over the strongest per-task baselines. PEARS also reduces the number of interaction episodes required for a certain success threshold by up to 53.2% relative to the fastest baseline. In real-world experiments, PEARS achieves success rates of 95% on Whiteboard Erasing and 90% on Pipette Liquid Aspiration. These results show that combining the PFR module with policy steering can accelerate adaptation while reducing costly interactions. The project website is available at https://song-kun.github.io/pears.

cs.RO↗
arXiv2026

No linear stretching of single-copy quantum pseudorandomness

Pseudorandom states, introduced by Ji, Liu, and Song (CRYPTO 2018), are quantum analogues of classical pseudorandom generators. A fundamental property of classical pseudorandom generators is that their output can be stretched to arbitrary polynomial length. Whether an analogous stretching property holds for quantum pseudorandom states has been elusive. Here, we essentially resolve this question for the single-copy variant of pseudorandom states, 1PRS. We show, via a unitary oracle separation, that no fully black-box construction can stretch the output length of a 1PRS by any constant factor greater than one.

quant-ph↗
arXiv2026

The Breakdown of Classical Minicrypt Equivalences in the Quantum-Computation Classical-Communication Model

Classically, it is well-known that several fundamental cryptographic primitives, including one-way functions, pseudorandom number generators, commitments, and signatures, characterize the same cryptographic world, which is known as "Minicrypt". In this paper, we investigate to what extent this picture persists in the quantum-computation classical-communication (QCCC) setting, where parties may perform quantum local computation, but communicate through only classical messages. We demonstrate that the classical Minicrypt landscape breaks down in the QCCC model by oracle separations. We construct two oracle worlds where efficiently verifiable one-way puzzles (EV-OWPuzz) exist. By a known equivalence, QCCC one-time signatures also exist in both worlds. In the first, one-way functions do not exist, even if we allow quantum pseudodeterministic evaluation. In the second, QCCC bit commitments do not exist. Thus, the classical Minicrypt reductions from signatures to one-way functions and bit commitments have no fully black-box counterparts in the QCCC setting. Our proofs develop techniques for analyzing the random phase states, including a concentration theorem and an LOCC decoupling theorem. As a separate result, we establish quantum sum-binding for the NOVY protocol, settling the question of its post-quantum security. We also generalize it to a fully black-box construction of a QCCC commitment from EV-OWPuzz with nearly uniform puzzle distribution, complementing our oracle separation results.

quant-ph↗
arXiv2026

Group-Marginalized Self-Rewarding RL Drives Zero-Label Self-Evolving

Self-rewarding reinforcement learning (RL) enables large language models (LLMs) to self-evolve without human labels. Existing ensemble-based methods construct reward references from rollout groups and assign rewards accordingly. However, a response's reward representation also depends on its randomly sampled group context, i.e., the other responses in its group. Using only one group-context realization may miss desired reward signals and provide unreliable guidance for policy optimization. To address this issue, we propose Group-Marginalized Advantage Estimation (GMAE), which aggregates reward realizations across possible contexts into a response-level distribution and estimates expected advantages. Experiments across eight benchmarks and four base models demonstrate strong performance and cross-domain generalization. GMAE also exhibits stable learning, low extra cost, and good applicability across training datasets and RL backbones.

cs.LG↗
arXiv2026

V-Engram: Trigger-Indexed External Memory for Modular Text-to-Image Personalization

Pretrained text-to-image models contain broad visual knowledge, yet they cannot reliably acquire or refine a specific visual identity from only a few references while preserving compositional control. Token-embedding methods are compact but often underfit identity, whereas adapter-based methods improve fidelity through persistent weight updates that can be costly to store and interfere when concepts are composed. We introduce V-Engram, a trigger-indexed external memory mechanism for Stable Diffusion 3.5. Each concept is assigned an explicit trigger that retrieves concept-specific memory, whose gated directions enter frozen text-encoder and MMDiT context states as relative residuals. Separating this memory from backbone adaptation enables prompt-selective and multi-concept access without merging model updates. Experiments show that V-Engram broadly matches DreamBooth-LoRA in overall subject fidelity while showing advantages in settings such as contextual subject preservation. Prompt-matched loading retrieves only matched entries, reducing most additional adaptation-state loading for a single-concept query. Qualitative results further demonstrate paired-trigger composition and same-class separation, while prompts without registered entries retain the frozen model's base behavior. Together, these results establish trigger-indexed memory as a modular interface for adding targeted visual evidence without rewriting the generator.

cs.AI↗
arXiv2026

GLIDE: Generalized Layer-wise Intrinsic Distributional Evaluation for Heterogeneous LLM Agents

LLM agents require reliable step-level evaluation to compare candidate branches and allocate computation effectively. However, lightweight evaluation remains challenging. External verifiers introduce additional inference cost, while agent-produced confidence or self-evaluation scores can be miscalibrated, especially when candidates are generated by heterogeneous agents. We propose \textbf{G}eneralized \textbf{L}ayer-wise \textbf{I}ntrinsic \textbf{D}istributional \textbf{E}valuation (\textbf{GLIDE}) for LLM agents. \textsc{GLIDE} derives intrinsic step evidence from layer-wise residual coherence, which measures whether local residual updates consistently support the global residual change induced by a candidate step. It calibrates this evidence against the recent score distribution of the generating agent and converts it into a pessimistic reward that jointly accounts for absolute residual evidence and agent-relative standing. The reward provides a cross-agent value signal for MCTS branch selection, while normalized predictive uncertainty guides adaptive branching. Experiments on multi-hop reasoning, sequential decision making, and symbolic logic show that \textsc{GLIDE} improves task performance, step-level ranking quality, and computational efficiency without external verifiers or task-specific supervision.

cs.AI↗
arXiv2026

Anon: Extrapolating Adaptivity Beyond SGD and Adam

Adaptive optimizers such as Adam and non-adaptive methods like SGD exhibit distinct generalization capabilities across different architectures. Prior tunable optimizers attempt to bridge this gap by strictly interpolating between SGD and Adam, effectively confining adaptivity within the 0-to-1 bound. However, this restricted interpolation is fundamentally insufficient: we reveal that optimal adaptivity often requires extrapolation, such as negative adaptivity for classical CNNs and adaptivity of at least one ($γ\geq 1$) for Transformers. Extrapolating adaptivity theoretically violates the strict non-decreasing pre-conditioner assumption, often leading to divergence in existing methods. To break this barrier, we propose Anon, an optimizer that achieves fully continuous adaptivity extrapolation across the entire real-number spectrum. To guarantee provable stability in these out-of-bound regimes, we introduce Incremental Delay Update (IDU), a novel mechanism that bypasses hard max-tracking strategies. We theoretically establish Anon's convergence in both convex and non-convex settings. Empirically, by exploring previously unreachable adaptivity landscapes, Anon demonstrates highly competitive and scalable performance among state-of-the-art element-wise optimizers on representative image classification, diffusion, and large language modeling tasks.

cs.AI↗
arXiv2026

ANO: Robust Policy Optimization via Bounded, Redescending Gain Fields

Proximal Policy Optimization (PPO) dominates reinforcement learning and LLM alignment, yet its hard-clipping mechanism and unconstrained alternatives (e.g., SPO) sit at two extremes of a stability-efficiency dilemma. We argue that this dilemma is best understood dynamically: a surrogate objective is a feedback law on the probability ratio, and its clipping/penalty shape defines a gain field that drives the update dynamics. PPO's clip induces a dead zone (zero feedback outside the trust region), leaving the policy to drift open-loop under momentum; SPO's quadratic penalty induces an unbounded, linearly growing gain that stiffens the dynamics and destabilizes under aggressive step sizes. Guided by this view, we derive Anchored Neighborhood Optimization (ANO), which designs the gain field directly: a $C^\infty$ shaping kernel that anchors the identity map at $r{=}1$, peaks exactly at a prescribed trust-region boundary $1{+}ε$, bounds the push on severely off-policy samples by a tunable $κ_{+}$, and exerts a bounded, redescending pull of tunable depth $κ_{-}$ on extreme outliers. The three hyperparameters have decoupled roles, and all internal constants are solved in closed form. Empirically, ANO ranks first on both Atari (40 games) and MuJoCo in IQM and Median of normalized scores. While the runner-up differs across domains (PAPO on Atari, SPO on MuJoCo), ANO is the only method consistently at the top. Under a learning-rate stress test ($3\times10^{-4}\!\to\!10^{-3}$), ANO degrades by only $0.9\%$ whereas PPO collapses by $54.5\%$, and the stressed ANO still outperforms PPO and PAPO at their best-tuned learning rates.

cs.AI↗
arXiv2026

GLoTouch: Global-to-Local Haptic Perception Using a Parallel Gripper for Object Search, Recognition, and Grasping Without External Vision

Perceiving objects in the environment is a fundamental capability of autonomous robots. In dark or low-light environments, external cameras often fail to reliably perceive object positions and geometry; when visual sensing is unavailable, completing target search, recognition, and grasping through touch alone becomes a key robot manipulation capability. This task must simultaneously address container-scale spatial exploration and object-scale fine-grained geometric perception, which is particularly challenging for low-degree-of-freedom parallel grippers. However, a unified framework remains lacking for connecting container-scale spatial exploration with object-scale fine-grained geometric perception and grasping. To address this challenge, we present \textbf{GLoTouch}, a global-to-local haptic perception and manipulation framework built on a parallel gripper. In the global stage, the gripper holds a passive long-reach probe, combining force measurements with known tool geometry to localize contacts and actively estimate candidate-object positions, coarse contours, and heights. In the local stage, the robot sets down the probe and uses the bilateral visuotactile sensors on the same gripper to directly acquire local haptic observations, which are matched against a given target 3-D model without object-specific training. We evaluate the framework in both simulation and real-robot experiments. Source code will be open-sourced.

cs.RO↗
arXiv2026

LLM can Read Spectrogram: Encoder-free Speech-Language Modeling

Recent speech-aware large language models (Speech-LLMs) rely on pre-trained speech encoders to convert audio into semantic/acoustic rich representations consumable by LLM. In this work, instead, we explore: can an LLM learn to read Mel spectrogram directly without a dedicated speech encoder? We propose Mel-LLM, an encoder-free Speech-LLM that feeds lightly pre-processed Mel-spectrogram patches directly into the LLM through a linear projection, allowing the LLM to learn speech-text alignment purely through its own parameters. We focus on speech understanding tasks, including automatic speech recognition (ASR), spoken QA and audio understanding. For ASR, we evaluate on the OpenASR Leaderboard public sets and production-level scaling experiments, demonstrating that the encoder-free solution achieves competitive performance with only limited degradation compared to encoder-initialized counterparts. We find that when data is limited, initialization from a multimodal checkpoint (Phi-4-MM) is crucial for maintaining performance. We also present ablation studies suggesting which LLM layers are most involved in speech adaptation. Beyond ASR, we extend Mel-LLM with general speech/audio understanding tasks, revealing an acoustic-semantic trade-off: directly exposing the LLM to Mel-spectrogram input improves paralinguistic and non-ASR acoustic tasks, while knowledge-intensive spoken QA remains more challenging than encoder-anchored systems. We additionally include a text-to-speech (TTS) proof-of-concept with a next-token VAE decoder, showing that direct Mel generation is possible but still trails stronger latent-diffusion generation.

eess.AS↗