arXiv subjects

Ji Wu

Publications and source records attributed to Ji Wu.

At least 19 recordsLinked to original sources

arXiv2026

AKDQ: Adaptive Key-Delta Quantization for Low-Bitrate RVQ-based Neural Speech Codecs

Low-bitrate speech coding is essential for bandwidth-constrained channels and on-device applications, but conventional residual vector quantization (RVQ)-based codecs employ a fixed number of codebooks per frame regardless of content. Adjacent speech frames often exhibit strong similarity, and quantizing them independently with the same scheme results in redundant encoding, thereby wasting bitrate. Therefore, we propose Adaptive Key-Delta Quantization (AKDQ), which replaces the fixed per-frame VQ allocation of RVQ with an adaptive frame-wise key-or-delta policy. AKDQ is a plug-in strategy that requires no modification to the codec's encoder-decoder backbone. Specifically, in AKDQ, each frame is classified as either a key-frame (K-frame) quantized with all VQs, or a delta-frame (D-frame) quantized with only a few VQs on a delta correction from the previous reconstruction, thereby reducing bitrate. The K/D decision is made by a budget-aware causal scheduler through a bit-bank mechanism, keeping the system streamable with no additional lookahead. Experiments on both causal and non-causal codec families demonstrate that AKDQ effectively reduces bitrate redundancy. For instance, on DAC, using AKDQ at 1.5 kbps achieves reconstruction quality comparable to using the baseline RVQ at 2.0 kbps, saving 25% of the bitrate.

eess.AS↗

arXiv2026

CRIM: Composable and Reusable Item Memory for Efficient Personalized Retrieval

Large language models incur substantial training and inference costs in personalized search and recommendation, as growing interaction histories lengthen request inputs and repeatedly expose the same item text to the model. We propose \textbf{C}omposable and \textbf{R}eusable \textbf{I}tem \textbf{M}emory (CRIM), a framework that separates item-level semantic encoding from request-level history modeling. A small set of learnable summary tokens compresses each item into compact continuous memories, while segment-level readout tokens (REC tokens) introduce predictive supervision within the interaction history. These memories are independent of users and queries and can therefore be reused across requests. At inference time, CRIM composes cached memories in interaction order and incorporates an optional query to produce a personalized retrieval representation, reducing repeated item-text encoding while retaining request-dependent contextual modeling. We further introduce a joint training mechanism that combines cached memories with sampled live re-encoding. Re-encoding only a few history items per request allows the retrieval objective to jointly optimize the item and request encoders at limited computational cost, while periodic cache refreshes maintain consistency. Extensive evaluation and cost analysis on MovieLens-1M, Qilin, and our new LazSeq dataset show that CRIM achieves state-of-the-art retrieval performance with substantially lower inference costs at the same backbone scale, while limiting item re-encoding during joint training. We will publicly release our code and dataset.

cs.IR↗

arXiv2026

Task-Level Natural Language Priors as Learning Signals for Low-Resource LLM Training

Large language models (LLMs) often struggle when low-resource training data are ambiguous or incomplete. Task-level natural-language priors can provide useful guidance in such settings, but existing approaches usually treat these priors as input context rather than as learning signals during training. We study how task-level natural-language priors can supply guidance missing from these data through Prior-Guided Tuning (PGT). Its implementation, Contrastive Prior Steering (CPS), retains the supervised objective and adds positive- and negative-prior-conditioned losses with opposing optimization directions. Across AmbiMath, Diagnostic Anchoring, and Jigsaw, CPS achieves near-perfect exact-match accuracy on the synthetic task, improves accuracy under misleading diagnostic suggestions by 8.2 percentage points, and improves average subgroup Macro F1 on Jigsaw by 10.7 percentage points over plain fine-tuning, averaged over the two evaluated models. On Jigsaw, with one tenth of the experimental training data, CPS slightly exceeds full-data plain fine-tuning. These results show how reusable task-level language can complement limited labeled examples without requiring instance-level rationales or preference pairs.

cs.AI↗

arXiv2026

Explainable and Generalisable LLM-based Cognitive Decline Detection with Spontaneous Speech

Alzheimer's disease (AD) and mild cognitive impairment (MCI), which may precede AD, manifest early through subtle linguistic and acoustic alterations. Traditional diagnostics, however, are often resource-intensive and lack scalability for mass screening. To address these challenges, we introduce a novel bilingual speech large language model framework for automated, explainable cognitive screening. Unlike conventional pipelines that rely on error-prone automatic speech recognition, our system directly processes raw speech to learn joint acoustic-semantic representations, preserving critical prosodic cues often lost in transcription. Utilising our newly collected PUTH-AD dataset alongside multiple open-source corpora, we implemented a multi-task learning objective that simultaneously performs cognitive status classification and generates clinician-understandable natural language explanations. Our system achieved the highest average accuracy and AUROC across six dataset/task conditions, comparing three representative baselines. The system demonstrated cross-task transfer to held-out PUTH-AD task subsets, maintaining classification accuracy on an entirely unseen cognitive task without task-specific fine-tuning. Furthermore, clinician evaluation confirms that the generated explanations are both clinically relevant and largely consistent with the underlying speech evidence, supporting their potential utility in clinical interpretation. This study provides a scalable, objective, and explainable framework for speech-based cognitive screening, combining cognitive status classification with natural language explanations that clinicians can assess and verify, bridging the gap between advanced AI and clinical utility.

eess.AS↗

arXiv2026

Magnetic Lifting of BIC Symmetry Protection Enables Nonreciprocal Beam Steering and Angular Information Encoding

Symmetry-protected bound states in the continuum (SP-BICs) offer exceptional optical confinement but remain nonradiative. Here, we demonstrate that magnetically lifting their protecting symmetry enables nonreciprocal beam steering with intrinsic angular information encoding. In a magneto-optical photonic crystal slab, an in-plane magnetic field converts an SP-BIC into a radiative quasi-BIC and, together with vertical dielectric asymmetry, continuously displaces the associated band extremum away from the Gamma point. The steered beam exhibits two independent locking relations: its wavelength is locked to the elevation angle, while its polarization orientation is locked to the azimuthal angle. These relations embed complementary spectral and polarization information into the scanning beam, providing additional sensing channels for multidimensional LiDAR. Moreover, the nonreciprocal dispersion gives different resonance frequencies at opposite wave vectors, supporting resonant emission at +k while back-reflected light at the same frequency is off resonance at -k, opening a route to intrinsic self-isolation and reduced optical feedback. These findings establish a unified framework connecting magnetic beam steering, angular information encoding, and nonreciprocal feedback suppression, with prospects for compact reconfigurable light sources and integrated optical engines.

physics.optics↗

arXiv2026

FitAQA: A Benchmark of Fitness Action Quality Assessment for Multimodal Large Language Models

Fitness Action Quality Assessment (AQA) is important for intelligent sports training, yet the capabilities of Multimodal Large Language Models (MLLMs) in this setting remain underexplored. Existing benchmarks rely on action-specific annotation schemes and focus primarily on final assessment outputs, offering limited insight into how models assess exercise quality. We introduce FitAQA, a systematic benchmark for evaluating MLLMs in fitness AQA, containing 2,219 videos and 5,512 QA instances across 30 bodyweight exercises. In collaboration with experts in sports science, we develop a unified form error taxonomy that defines 38 recurring form errors within six complementary quality dimensions: alignment, symmetry, stability, coordination, tempo, and completeness. This taxonomy provides a shared assessment framework across different exercises. FitAQA further formulates three evaluation tasks: perception for recognizing relevant visual evidence, judgement for combining that evidence with domain knowledge to assess execution correctness, and temporal grounding for localizing form errors over time. Extensive evaluation shows that current MLLMs still struggle to assess exercise quality comprehensively and localize form errors precisely. Controlled experiments further indicate that visual perception is a key bottleneck, as judgement performance improves substantially when ground-truth perceptual evidence is provided. The dataset is available at https://huggingface.co/datasets/Kelly0510/FitAQA.

cs.AI↗

arXiv2026

Relationally Guided Use Case Modeling with LLMs

Use case flows are important elements of use case modeling because they support downstream software engineering activities, including requirements analysis, architectural and detailed design, and test case generation. However, constructing them manually is costly and expertise-intensive, while existing automated approaches still struggle to preserve semantic consistency, control-flow logic, data-flow logic, and the intended system boundary, especially when identifying branch points and generating alternative flows. To address this problem, we propose FlowGen for complete use case flow construction. FlowGen uses LLM-based Semantic Information Processing (SIP) to extract semantic elements, constructs a Semantic Relational Graph (SRG) encoded by an enhanced R-GAT for basic flow generation (BFGen), and further supports branch point prediction through BPP and branch-conditioned alternative flow generation through AFGen. Evaluations on 13 public and 7 industrial datasets show that FlowGen consistently outperforms competitive baselines in all three core components. In particular, BFGen improves over the best baseline by 14% in Precision, 7-25% in Recall, 11-30% in F1, and 10-19% in AUC; BPP improves Precision by 30-110%, Recall by 33-91%, and F1 by 32-117%; AFGen improves Precision by 8-23%, F1 by 5-18%, and AUC by 0.6-2.5%. Moreover, we validate the effectiveness of the LLM-based SIP module and the attention preservation factor in BFGen, analyze the impact of requirement completeness on BFGen, and examine how different scopes of branch-related context affect AFGen.

cs.SE↗

arXiv2026

Pose2Muscle: Structured Spatio-Temporal Decoding for Discrete Muscle Activity Estimation from Human Pose

Muscle activity is fundamental to human movement, and understanding its patterns is critical for injury prevention and rehabilitation. Conventional muscle activity monitoring relies on specialized sensors such as surface electromyography, which limits its practicality for long-term real-world use. Existing studies suggest that muscle-related information can be inferred from human pose. However, the substantial gap between externally observable pose and internal muscle activation, limits the accuracy and generalization of current approaches. In this study, we propose Pose2Muscle, a pose-driven framework for discrete muscle activity estimation without requiring sEMG signals at inference time. Instead of directly regressing continuous sEMG signals, Pose2Muscle reformulates muscle estimation as a structured prediction problem over discrete muscle activity states, yielding a more stable and interpretable target space. The framework combines multi-scale spatio-temporal attention to capture motion patterns at complementary spatial and temporal scales with a directed acyclic graph-based decoder that maintains multiple candidate muscle-state hypotheses and performs structured trajectory inference over time. To support this task, we construct PoseEMG-43, a synchronized pose-sEMG dataset containing 2,992 movement instances from 43 daily-life actions performed by 14 participants. Experiments show that Pose2Muscle consistently outperforms representative retrieval- and pose-based baselines. It achieves an Adjacent-level Accuracy of 86.36% and a Pearson correlation coefficient of 0.8821 under the Random Split, and 63.97% and 0.6795, respectively, under the Subject-Level Split. These results demonstrate the feasibility of inferring structured muscle-state patterns from human pose and suggest the potential of Pose2Muscle for muscle-aware movement analysis when direct physiological sensing is impractical

cs.CV↗

arXiv2026

From Continuous sEMG Signals to Discrete Muscle State Tokens: A Robust and Interpretable Representation Framework

Surface electromyography (sEMG) signals exhibit substantial inter-subject variability and are highly susceptible to noise, posing challenges for robust and interpretable decoding. To address these limitations, we propose a discrete representation of sEMG signals based on a physiology-informed tokenization framework. The method employs a sliding window aligned with the minimal muscle contraction cycle to isolate individual muscle activation events. From each window, ten time-frequency features, including root mean square (RMS) and median frequency (MDF), are extracted, and K-means clustering is applied to group segments into representative muscle-state tokens. We also introduce a large-scale benchmark dataset, ActionEMG-43, comprising 43 diverse actions and sEMG recordings from 16 major muscle groups across the body. Based on this dataset, we conduct extensive evaluations to assess the inter-subject consistency, representation capacity, and interpretability of the proposed sEMG tokens. Our results show that the token representation exhibits high inter-subject consistency (Cohen's Kappa = 0.82+-0.09), indicating that the learned tokens capture consistent and subject-independent muscle activation patterns. In action recognition tasks, models using sEMG tokens achieve Top-1 accuracies of 75.5% with ViT and 67.9% with SVM, outperforming raw-signal baselines (72.8% and 64.4%, respectively), despite a 96% reduction in input dimensionality. In movement quality assessment, the tokens intuitively reveal patterns of muscle underactivation and compensatory activation, offering interpretable insights into neuromuscular control. Together, these findings highlight the effectiveness of tokenized sEMG representations as a compact, generalizable, and physiologically meaningful feature space for applications in rehabilitation, human-machine interaction, and motor function analysis.

eess.SP↗

arXiv2026

FAMPWQ: Fisher Information-based Adaptive Mixed Precision Weight Quantization for Effective LLM Inference

Recent years have witnessed remarkable achievements of Large Language Models (LLMs) in multiple domains, while the excessive resource requirements of LLMs hinder the deployment on resource-constrained devices. Although model quantization stands out as an effective approach, conventional quantization approaches typically incur severe performance degradation due to uniform bit-width or simple heuristic sensitivity evaluation. In this paper, we propose a novel Fisher information-based Adaptive Mixed Precision Weight Quantization approach, i.e., FAMPWQ, which performs layer-adaptive weight quantization for effective LLM inference on commodity GPUs. First, we propose a system model with a novel Fisher information metric to measure the layer-wise sensitivity to quantization. Second, we propose a reinforcement learning-based bit-width allocator in FAMPWQ, which generates an adaptive bit-width allocation strategy based on the Fisher information sensitivity metric. Extensive experiments on 7 models and 5 benchmarks demonstrate that FAMPWQ significantly outperforms 7 baseline approaches in terms of PPL (up to 3.39 smaller), accuracy (up to 6.87% higher), and LLM-as-a-judge comparison (up to 76% win rate).

cs.LG↗

arXiv2026

Quantifying Entrainment Evidence: A Comparison of Frequentist and Bayesian Approaches for Information Processing Pathway Maps

Information Processing Pathway Maps (IPPMs) offer a scalable framework for formalizing the complex sequence of mathematical transformations applied to sensory stimuli. These maps chart the latency and cortical expression of computational steps, relying on statistical inference to link model outputs with observed neural activity. Traditionally, this mapping has relied on frequentist hypothesis testing. However, determining which of several competing computational models best explains neural data is a problem of model adjudication, arguably better suited to probabilistic inference. Here, we present a direct comparison between the established frequentist approach and a novel Bayesian framework for mapping cortical entrainment. While the Bayesian formulation retains the core strength of IPPMs -- generating explicit predictions of time-varying neural signals -- it fundamentally alters the selection criterion, shifting from rejecting a null hypothesis to quantifying the relative evidence for competing computational hypotheses. We evaluate the performance and interpretability of both approaches using an auditory neuroimaging dataset to reconstruct a known loudness-processing pathway. We discuss the implications of this shift for systems neuroscience, specifically regarding the handling of collinear models and the robust accumulation of evidence.

q-bio.NC↗

arXiv2026

HOBA: Hierarchical On-Policy Bidding Agents for Adaptive Online Advertising

Online advertising bidding systems typically deploy multiple offline-trained expert models (e.g., PID controllers, model predictive control, offline RL policies) but face two critical limitations: lack of online adaptability to non-stationary auction markets, and reliance on costly manual tuning of hyperparameters such as bid bounds and budget pacing constraints. We propose HOBA (Hierarchical On-policy Bidding Agents), a hierarchical reinforcement learning framework that decouples strategic reasoning, model selection, and bid execution across three time scales. At the high level, a large language model infers hyperparameters from contextual signals through a Think-Act-Observe-Reflect loop with historical experience retrieval. At the mid level, a SARSA agent dynamically selects among expert models, incorporating causal adjustment to eliminate selection bias. At the low level, a dynamic expert pool (PID, MPC, IQL, Decision Transformer) executes bids under high-level constraints. This design confines online learning to discrete expert selection rather than continuous bid optimization, significantly reducing exploration risk while maintaining adaptability. Experiments on the AuctionNet benchmark and a large-scale A/B test demonstrate consistent improvements over state-of-the-art baselines. In a large-scale online deployment, HOBA delivered substantial business value, achieving a +3.6\% increase in target cost, proving the effectiveness of our hierarchical multi-agent bidding paradigm.

cs.AI↗

arXiv2026

ReportQA: QA-Based Radiology Report Evaluation

Radiology report evaluation is essential for advancing automated report generation. Natural language generation metrics have limited clinical relevance. Clinical efficacy (CE) metrics evaluate important medical findings, but focus mainly on presence and cover only a limited set of entities. Due to heavy reliance on manual annotations, it is difficult for CE metrics to extend clinical entities or attributes. In clinical practice, radiology reports serve as a medium for information transfer. Clinicians use them to perform downstream diagnostic tasks without directly inspecting images. Based on this insight, we propose ReportQA, a clinical-related and flexible radiology report evaluation framework, supporting detailed quantitative analysis of radiology report generation systems. We first collect datasets covering multiple imaging modalities and anatomical regions. We then construct knowledge trees of clinical entities and attributes with radiologist guidance, and use large language models (LLMs) to extract structured information from raw reports. Next, we generate QA pairs from predefined templates and apply quality control through self-filtering and report-based filtering. During evaluation, the report is treated as context, and an LLM acts as a judge model to answer the QA pairs. Based on the resulting QA accuracy, we introduce QAScore metric. Compared with existing metrics, QAScore shows better alignment with radiologist judgments. Experiments on multiple state-of-the-art vision-language models reveal that current report-based inference paradigms struggle to learn fine-grained clinical representations and exhibit strong negative prior biases. In contrast, question-driven inference provides a more effective alternative. For reproducibility and extensibility, we release the knowledge trees, structured reports, and QA pairs, along with the pipeline code for QA construction and evaluation.

cs.CL↗

arXiv2026

The 1st PortraitCraft Challenge: A CVPR 2026 Workshop Competition on Portrait Composition Understanding and Generation

This paper presents an overview of the inaugural PortraitCraft Challenge, held as one of the official competitions at CVPR 2026. The challenge focuses on portrait composition understanding and generation, aiming to advance AI research in portrait aesthetics analysis and controllable image synthesis. Unlike existing datasets and tasks that primarily focus on global aesthetic scoring, PortraitCraft introduces a unified evaluation framework comprising two complementary tracks. Track 1 requires models to perform structured portrait composition understanding, and Track 2 requires models to generate portrait images from structured composition descriptions under explicit compositional constraints. To support the challenge, we constructed and publicly released a large-scale portrait composition dataset consisting of approximately 50,000 curated real portrait images, providing multi-level supervision. This report describes the challenge setup, evaluation protocols, dataset composition, and final results, along with an analysis of the technical characteristics of the submitted solutions. The PortraitCraft Challenge provides a standardized and reproducible platform for research on portrait composition understanding and generation, and is expected to foster further progress in the fields of portrait aesthetics and controllable image generation.

cs.CV↗

arXiv2026

An Ultra-Low-Bitrate Neural Speech Codec with Plain-to-Pseudo Synergistic Vector Quantization

Most neural speech codecs use residual vector quantization (RVQ), in which later VQs contribute less but consume the same bitrate, leading to inefficiency. We propose P2PSynCodec, an ultra-low-bitrate neural speech codec with a plain-to-pseudo synergistic vector quantizer (P2PSVQ). P2PSVQ consists of one plain VQ and multiple pseudo VQs. The plain VQ produces basic tokens by quantization, while the pseudo VQs generate auxiliary tokens by neural prediction and incur zero transmitted bitrate. Thus, speech is decoded from the plain-VQ tokens together with predicted pseudo-VQ tokens, greatly reducing bitrate. Experiments show that P2PSynCodec achieves speech reconstruction quality comparable to competing codecs at 2.0 kbps while operating at only 0.5 kbps, demonstrating high efficiency for ultra-low-bitrate speech coding.

eess.AS↗

arXiv2026

VoCodec: A Low-bitrate Streamable Neural Speech Codec with Voicing-driven Quantization

Neural speech codecs are key to speech transmission and storage, but most use uniform quantization across frames, allocating the same bitrate regardless of content and wasting bits. We propose VoCodec, a low-bitrate streamable neural speech codec with voicing-driven quantization that assigns higher bitrate to voiced frames and lower bitrate to unvoiced frames according to perceptual sensitivity. VoCodec embeds a voicing detector in a fully causal encoder-quantizer-decoder neural coding framework, using residual scalar-vector quantization for voiced frames and simple scalar quantization for unvoiced ones. Experiments show that on the LibriTTS dataset at a 16 kHz sampling rate, VoCodec outperforms baseline neural speech codecs even at a bitrate as low as 1.1 kbps. Our further experiments also confirm that introducing voicing-driven quantization can effectively reduce the bitrate by approximately 27% compared with uniform quantization strategy.

eess.AS↗

arXiv2026

CFMDCTCodec: A Low-Bitrate Neural Speech Codec with Noise-Prior-aware Conditional Flow Matching for MDCT-Spectral Enhancement

High-quality speech coding at low bitrates is crucial for bandwidth-constrained applications, yet remains challenging due to the severe loss of quality-critical information in highly compressed representations. To overcome this challenge, we propose CFMDCTCodec, a low-bitrate neural speech codec that operates entirely in the modified discrete cosine transform (MDCT) domain. CFMDCTCodec integrates a lightweight encoder-quantizer-decoder-style MDCT-spectral codec with a noise-prior-aware, conditional-flow-matching (CFM)-based MDCT-spectral enhancer. Within this framework, the codec serves as a base module that compactly discretizes the MDCT spectrum extracted from speech and produces an initial coarse reconstruction, while the enhancer further restores fine-grained spectral details. The enhancer improves the decoded MDCT spectrum by integrating a conditional MDCT velocity-field filter with an ordinary differential equation (ODE) solver, under the guidance of an MDCT-derived magnitude-adaptive noise prior, aiming to emphasize perceptually significant high-energy regions while stabilizing low-energy and silent regions. Finally, the enhanced MDCT spectrum is reconstructed into the decoded speech using the inverse MDCT. When optimizing CFMDCTCodec, we adopt a unified non-adversarial training strategy that jointly combines reconstruction, quantization and CFM objectives. Both objective and subjective evaluations show that CFMDCTCodec outperforms competitive baselines in low-bitrate regimes, e.g., 0.65 kbps, while approaching the perceptual quality of large-scale codecs with significantly fewer parameters and computations.

eess.AS↗

arXiv2026

InterMesh: Explicit Interaction-Aware End-to-End Multi-Person Human Mesh Recovery

Humans constantly interact with their surroundings. Existing end-to-end multi-person human mesh recovery methods, typically based on the DETR framework, capture inter-human relationships through self-attention across all human queries. However, these approaches model interactions only implicitly and lack explicit reasoning about how humans interact with objects and with each other. In this paper, we propose InterMesh, a simple yet effective framework that explicitly incorporates human-environment interaction information into human mesh recovery pipeline. By leveraging a human-object interaction detector, InterMesh enriches query representations with structured interaction semantics, enabling more accurate pose and shape estimation. We design lightweight modules, Contextual Interaction Encoder and Interaction-Guided Refiner, to integrate these features into existing HMR architectures with minimal overhead. We validate our approach through extensive experiments on 3DPW, MuPoTS, CMU Panoptic, Hi4D, and CHI3D datasets, demonstrating remarkable improvements over state-of-the-art methods. Notably, InterMesh reduces MPJPE by 9.9% on CMU Panoptic and 8.2% on Hi4D, highlighting its effectiveness in scenarios with complex human-object and inter-human interactions. Code and models are released at https://github.com/Kelly510/InterMesh.

cs.CV↗