arXiv subjects

Li Su

Publications and source records attributed to Li Su.

At least 19 recordsLinked to original sources

arXiv2026

VioTech: Frame-Level Violin Playing Technique Detection with Multi-Scale Mixture-of-Experts

Playing technique is crucial for expressive violin performance, yet few violin technique detection models exist. Current systems mainly rely on note-level pipelines that first estimate note boundaries and then classify techniques. This is especially problematic for bowed string instruments such as the violin, where soft onsets and continuous bow-string interaction render note segmentation unreliable. Here, we propose VioTech, a frame-level violin technique detection framework that predicts techniques directly at 10 ms resolution without explicit note segmentation. To handle the large temporal-scale variation among techniques, VioTech combines a frozen MERT encoder with a multi-scale Mixture-of-Experts (MoE) architecture and task-specific routing for tonal, articulation, and legato-related predictions. We also introduce VioTech120, the first manually annotated dataset of violin playing techniques constructed from real solo performance recordings, as well as an accompanying note-labeling user interface. Results demonstrate that VioTech achieved consistent improvements over prior methods and significantly enhanced cross-dataset generalization, establishing a new state-of-the-art in violin technique-aware transcription.

cs.SD↗

arXiv2026

DJustify: Reason-and-Verify DJ Planning for Pop Transitions and Story-Driven Sets

Automatic DJ systems mostly target electronic dance music, whose regular structure makes rule-based mixing relatively easier. Vocal-rich popular music is harder: which track to play next, where to cue it, and which transition to use are tied together, so fixed rules handle them poorly. We present DJustify, a pop AI-DJ that puts an LLM in a reason--propose--verify--repair loop and lets it choose the next track, cue points, and mixing technique together. The LLM samples several transition plans; an external acoustic checker then scores the rendered audio and triggers repair when a check fails. Sampling keeps several candidate transitions per track pair instead of one fixed mix. In blind listening tests our transitions beat Mosaikbox, a reimplemented crossfade, and a cue baseline that reuses the crossfade's cues; that last comparison shows most of the margin comes from cue placement. The same planner also picks and orders a set along an emotion arc: this improves story coherence over a no-story ablation, and within a set story and transition ratings show no measurable gap from a human DJ. The checker mainly adds reliability, not higher average quality: it repairs rare but audible errors such as cut vocals, while the planner stays non-deterministic.

cs.SD↗

arXiv2026

An Optical Music Recognition Tool for String Quartets

This paper presents an optical music recognition (OMR) system tailored for PDF scores of string quartet pieces. The approach begins with a SegNet-based segmentation model to localize musical regions, followed by convolutional neural networks and image processing techniques to classify and extract musical symbols. To improve rhythmic precision, dynamic programming methods enforce time signature constraints and explicitly handle tuplets. Evaluation using a longest common subsequence metric across multiple feature matrices demonstrates robust accuracy in recognizing pitch, onset, duration, and accidentals on a benchmark dataset. Code and data are available at https://github.com/lattellie/25-omr.

cs.CV↗

arXiv2026

Pitch Smoothing Using Relative Interval Networks

Pitch tracking systems typically couple a per-frame fundamental frequency ($F_0$) estimator with a temporal smoothing stage to obtain continuous trajectories. Conventional Viterbi smoothers enforce first-order continuity but lack long-term temporal awareness and could lock into octave errors across corrupted frames. We propose Relative Interval Networks (RIN), a trajectory smoothing framework that reconciles per-frame pitch estimates with data-driven multi-hop pitch differences. We extract robust relative pitch intervals across arbitrary frame offsets using Variable-Q Transform cross-correlation. We formulate pitch smoothing as an $L_1$-norm optimization problem and prove its equivalence to a minimum cost circulation problem, solved efficiently via linear programming. Evaluations across speech, singing, and instrumental datasets show that RIN substantially improves weak estimators, matches or outperforms Viterbi decoding at a comparable computational cost, and provides superior robustness under certain acoustic degradation.

eess.AS↗

arXiv2026

Separate-and-Detect: Unified Drum Transcription and Stem Generation via Latent Diffusion

Automatic Drum Transcription (ADT) is commonly formulated as a direct mapping from a music mixture to symbolic drum events. While effective for transcription, this formulation discards the acoustic stems that are useful for editing, remixing, and production. We revisit an alternative separate-and-detect formulation, where a drum source separation front end first produces five editable drum stems, and a fixed onset detector then converts each stem into symbolic events. The separator is built on a five-stem latent diffusion model that jointly generates kick, snare, toms, hi-hats, and cymbals in a compact VAE latent space. We further study two training-only auxiliary branches--an onset branch (OB) and a timbre branch (TB)--which shape the separator during learning but are discarded at inference. Trained on synthetic drum multitracks and evaluated on MDB Drums and ENST-Drums, the proposed pipeline consistently improves over a strong U-Net-based drum separation baseline in overall transcription F1. It also outperforms a representative end-to-end ADT system on kick and snare F1 under our evaluation protocol, while additionally providing separated audio stems. The ablation results show that OB gives the most stable transcription gains, whereas TB changes the trade-off between reconstruction, acoustic stem quality, and onset detection. These results suggest that generative drum demixing can serve not only as a source separation model, but also as a practical front end for interpretable drum transcription.

cs.SD↗

arXiv2026

Improving Longitudinal Targeted Maximum Likelihood Estimation in Target Trial Emulation using Joint Calibrated Weights

In target trial emulation (TTE), marginal structural models (MSMs) can be used to characterise per-protocol treatment effects over time. The MSM parameters are often estimated by inverse probability weighting (IPW), with weights estimated by maximum likelihood. However, IPW-based estimators can be unstable in small samples and are sensitive to misspecification of the weight models. An alternative method for estimating the MSM parameters is longitudinal targeted maximum likelihood estimation (LTMLE). LTMLE is double robust and potentially more efficient than IPW. Nevertheless, LTMLE also relies on inverse probability weights and may therefore share the instability of IPW-based estimators. We propose joint calibrated LTMLE, which integrates LTMLE with joint calibrated weights tailored for per-protocol effect estimation in TTE. This calibration of weights improves finite-sample performance by enforcing covariate balance in both the treatment and censoring processes simultaneously. Simulations show that the proposed method has improved efficiency and robustness to weight model misspecification, compared to standard LTMLE. We illustrate the method using a case study to evaluate the effect of highly active antiretroviral therapy on CD4 cell count among HIV-positive women.

stat.ME↗

arXiv2026

Spectral extremal problems for the $(p,Q)$-spectral radius of hypergraphs

Let $Q$ be an $s$-vertex $r$-uniform hypergraph, and let $H$ be an $n$-vertex $r$-uniform hypergraph. Denote by $\mathcal{N}(Q,H)$ the number of isomorphic copies of $Q$ in $H$. For a hereditary family $\mathcal{P}$ of $r$-uniform hypergraphs, define $$π(Q,\mathcal{P}):=\lim\limits_{n\to \infty}\binom{n}{s}^{-1}\max\{\mathcal{N}(Q,H): H\in \mathcal{P}~~\mbox{and}~~|V(H)|=n\}.$$ For $p\geq1$, the $(p,Q)$-spectral radius of $H$ is defined as $$λ^{(p)}(Q,H):=\max_{\|\mathbf{x}\|_{p}=1}s!\sum_{\{i_{1},\ldots,i_{s}\}\in \binom{[n]}{s}}\mathcal{N}(Q,H[\{i_{1},\ldots,i_{s}\}])x_{i_{1}}\cdots x_{i_{s}}.$$ In this paper, we present a systematically investigation of the parameter $λ^{(p)}(Q,H)$. First, we prove that the limit $$λ^{(p)}(Q,\mathcal{P}):=\lim\limits_{n\to \infty}n^{s/p-s}\max\{λ^{(p)}(Q,H): H\in \mathcal{P}~~\mbox{and}~~|V(H)|=n\}$$ exists, and for $p>1$, it satisfies $$π(Q,\mathcal{P})=λ^{(p)}(Q,\mathcal{P}).$$ Second, we study spectral generalized Turán problems. Specifically, we establish a spectral stability result and apply it to derive a spectral version of the Erdős Pentagon Problem: for $p\geq1$ and sufficiently large $n$, the balanced blow-up of $C_{5}$ maximizes $λ^{(p)}(C_{5},H)$ among all $n$-vertex triangle-free graphs $H$, thereby improving a result of Liu \cite{Liu2025}. Furthermore, we show that for $p\geq1$ and sufficiently large $n$, the $l$-partite Turán graph $T_{l}(n)$ attains the maximum $λ^{(p)}(K_{s},H)$ among all $n$-vertex F-free graphs $H$, where $F$ is an edge-critical graph with $χ(F)=l+1$. This provides a spectral analogue of a theorem due to Ma and Qiu \cite{MQ2020}.

math.CO↗

arXiv2026

Spectral Turán-type problems for the $α$-spectral radius of hypergraphs with degree stability

An $r$-pattern $P$ is an ordered pair $P=([l],E)$, where $l$ is a positive integer and $E$ is a set of $r$-multisets with elements from $[l]$. An $r$-graph $H$ is said to be $P$-colorable if there is a homomorphism $ϕ$: $V(H)\rightarrow [l]$ such that $\{ϕ(v_{1}),\ldots,ϕ(v_{r})\}\in E$ for every edge $\{v_{1},\ldots,v_{r}\}\in E(H)$. Let $\mathrm{Col}(P)$ denote the family of all $P$-colorable $r$-graphs. This paper studies spectral extremal problems for $α$-spectral radius of hypergraphs via analytic techniques. We first prove that for any $r$-pattern $P$, the hypergraph attaining the maximum $α$-spectral radius in $\mathrm{Col}(P)$ is asymptotically regular. Specifically, we establish asymptotically tight lower bounds for the minimum component of the principal eigenvector and the minimum degree of the spectral extremal hypergraphs in $\mathrm{Col}(P)$. Building on this regularity, we further show that for any family $\mathcal{F}$ of $r$-graphs that is degree-stable with respect to $\mathrm{Col}(P)$, spectral Turán-type problems can be completely reduced to spectral extremal problems within $\mathrm{Col}(P)$. As an application, we determine the maximum $α$-spectral radius ($α\geq1$) among all $n$-vertex $F^{(r)}$-free $r$-graphs, where $F^{(r)}$ is the $r$-expansion of the color-critical graph $F$. This provides a powerful reduction tool for handling spectral Turán-type problems in hypergraphs. Finally, leveraging the spectral method, we derive a corresponding edge Turán extremal result. More precisely, we show that if $\mathcal{F}$ is degree-stable with respect to $\mathrm{Col}(P)$, then every $\mathcal{F}$-free edge extremal hypergraph must be a $P$-colorable hypergraph.

math.CO↗

arXiv2026

Probability-Invariant Random Walk Learning on Gyral Folding-Based Cortical Similarity Networks for Alzheimer's and Lewy Body Dementia Diagnosis

Alzheimer's disease (AD) and Lewy body dementia (LBD) present overlapping clinical features yet require distinct diagnostic strategies. While neuroimaging-based brain network analysis is promising, atlas-based representations may obscure individualized anatomy. Gyral folding-based networks using three-hinge gyri provide a biologically grounded alternative, but inter-individual variability in cortical folding results in inconsistent landmark correspondence and highly irregular network sizes, violating the fixed-topology and node-alignment assumptions of most existing graph learning methods, particularly in clinical datasets where pathological changes further amplify anatomical heterogeneity. We therefore propose a probability-invariant random-walk-based framework that classifies individualized gyral folding networks without explicit node alignment. Cortical similarity networks are built from local morphometric features and represented by distributions of anonymized random walks, with an anatomy-aware encoding that preserves permutation invariance. Experiments on a large clinical cohort of AD and LBD subjects show consistent improvements over existing gyral folding and atlas-based models, demonstrating robustness and potential for dementia diagnosis.

q-bio.NC↗

arXiv2026

VioPTT: Violin Technique-Aware Transcription from Synthetic Data Augmentation

While automatic music transcription is well-established in music information retrieval, most models are limited to transcribing pitch and timing information from audio, and thus omit crucial expressive and instrument-specific nuances. One example is playing technique on the violin, which affords its distinct palette of timbres for maximal emotional impact. Here, we propose VioPTT (Violin Playing Technique-aware Transcription), a lightweight cascade model that directly transcribes violin playing technique in addition to pitch onset and offset. Furthermore, we release MOSA-VPT, a novel, high-quality synthetic violin playing technique dataset to circumvent the need for manually labeled annotations. Leveraging this dataset, our model demonstrated strong generalization to real-world note-level violin technique recordings in addition to achieving state-of-the-art transcription performance. To our knowledge, VioPTT is the first to jointly combine violin transcription and playing technique prediction within a unified framework.

cs.SD↗

arXiv2026

SynthCloner: Synthesizer-style Audio Transfer via Factorized Codec with ADSR Envelope Control

Electronic synthesizer sounds are controlled by parameter settings that yield complex timbral characteristics and ADSR envelopes, making synthesizer-style audio transfer particularly challenging. Recent approaches to timbre transfer often rely on spectral objectives or implicit style matching, offering limited control over envelope shaping. Moreover, public synthesizer datasets rarely provide diverse coverage of timbres and ADSR envelopes. To address these gaps, we present SynthCloner, a factorized codec model that disentangles audio into three attributes: ADSR envelope, timbre, and content. This separation enables expressive audio transfer with independent control over these attributes. Additionally, we introduce SynthCAT, a new synthesizer dataset with a task-specific rendering pipeline covering 250 timbres, 120 ADSR envelopes, and 100 MIDI sequences. Experiments show that SynthCloner outperforms baselines on both objective and subjective metrics, while enabling independent attribute control. The code, model checkpoint, and audio examples are available at https://buffett0323.github.io/synthcloner/.

eess.AS↗

arXiv2026

STaR: Sensitive Trajectory Regulation for Unlearning in Large Reasoning Models

Large Reasoning Models (LRMs) have advanced automated multi-step reasoning, but their ability to generate complex Chain-of-Thought (CoT) trajectories introduces severe privacy risks, as sensitive information may be deeply embedded throughout the reasoning process. Existing Large Language Models (LLMs) unlearning approaches that typically focus on modifying only final answers are insufficient for LRMs, as they fail to remove sensitive content from intermediate steps, leading to persistent privacy leakage and degraded security. To address these challenges, we propose Sensitive Trajectory Regulation (STaR), a parameter-free, inference-time unlearning framework that achieves robust privacy protection throughout the reasoning process. Specifically, we first identify sensitive content via semantic-aware detection. Then, we inject global safety constraints through secure prompt prefix. Next, we perform trajectory-aware suppression to dynamically block sensitive content across the entire reasoning chain. Finally, we apply token-level adaptive filtering to prevent both exact and paraphrased sensitive tokens during generation. Furthermore, to overcome the inadequacies of existing evaluation protocols, we introduce two metrics: Multi-Decoding Consistency Assessment (MCS), which measures the consistency of unlearning across diverse decoding strategies, and Multi-Granularity Membership Inference Attack (MIA) Evaluation, which quantifies privacy protection at both answer and reasoning-chain levels. Experiments on the R-TOFU benchmark demonstrate that STaR achieves comprehensive and stable unlearning with minimal utility loss, setting a new standard for privacy-preserving reasoning in LRMs.

cs.AI↗

arXiv2026

Timed text extraction from Taiwanese Kua-á-hì TV series

Taiwanese opera (Kua-á-hì), a major form of local theatrical tradition, underwent extensive television adaptation notably by pioneers like Iûnn Lē-hua. These videos, while potentially valuable for in-depth studies of Taiwanese opera, often have low quality and require substantial manual effort during data preparation. To streamline this process, we developed an interactive system for real-time OCR correction and a two-step approach integrating OCR-driven segmentation with Speech and Music Activity Detection (SMAD) to efficiently identify vocal segments from archival episodes with high precision. The resulting dataset, consisting of vocal segments and corresponding lyrics, can potentially supports various MIR tasks such as lyrics identification and tune retrieval. Code is available at https://github.com/z-huang/ocr-subtitle-editor .

cs.SD↗

arXiv2025

Erkang-Diagnosis-1.1 Technical Report

This report provides a detailed introduction to Erkang-Diagnosis-1.1 model, our AI healthcare consulting assistant developed using Alibaba Qwen-3 model. The Erkang model integrates approximately 500GB of high-quality structured medical knowledge, employing a hybrid approach combining enhanced pre-training and retrieval-enhanced generation to create a secure, reliable, and professional AI health advisor. Through 3-5 efficient interaction rounds, Erkang Diagnosis can accurately understand user symptoms, conduct preliminary analysis, and provide valuable diagnostic suggestions and health guidance. Designed to become users intelligent health companions, it empowers primary healthcare and health management. To validate, Erkang-Diagnosis-1.1 leads GPT-4 in terms of comprehensive medical exams.

cs.AI↗

arXiv2025

LargeSHS: A large-scale dataset of music adaptation

Recent advances in AI-based music generation have focused heavily on text-conditioned models, with less attention given to reference-based generation such as song adaptation. To support this line of research, we introduce LargeSHS, a large-scale dataset derived from SecondHandSongs, containing over 1.7 million metadata entries and approximately 900k publicly accessible audio links. Unlike existing datasets, LargeSHS includes structured adaptation relationships between musical works, enabling the construction of adaptation trees and performance clusters that represent cover song families. We provide comprehensive statistics and comparisons with existing datasets, highlighting the unique scale and richness of LargeSHS. This dataset paves the way for new research in cover song generation, reference-based music generation, and adaptation-aware MIR tasks.

cs.SD↗

arXiv2025

Enhancing Automatic Chord Recognition through LLM Chain-of-Thought Reasoning

Music Information Retrieval (MIR) encompasses a broad range of computational techniques for analyzing and understanding musical content, with recent deep learning advances driving substantial improvements. Building upon these advances, this paper explores how large language models (LLMs) can serve as an integrative bridge to connect and integrate information from multiple MIR tools, with a focus on enhancing automatic chord recognition performance. We present a novel approach that positions text-based LLMs as intelligent coordinators that process and integrate outputs from diverse state-of-the-art MIR tools-including music source separation, key detection, chord recognition, and beat tracking. Our method converts audio-derived musical information into textual representations, enabling LLMs to perform reasoning and correction specifically for chord recognition tasks. We design a 5-stage chain-of-thought framework that allows GPT-4o to systematically analyze, compare, and refine chord recognition results by leveraging music-theoretical knowledge to integrate information across different MIR components. Experimental evaluation on three datasets demonstrates consistent improvements across multiple evaluation metrics, with overall accuracy gains of 1-2.77% on the MIREX metric. Our findings demonstrate that LLMs can effectively function as integrative bridges in MIR pipelines, opening new directions for multi-tool coordination in music information retrieval tasks.

cs.SD↗

arXiv2025

Is Transfer Learning Necessary for Violin Transcription?

Automatic music transcription (AMT) has achieved remarkable progress for instruments such as the piano, largely due to the availability of large-scale, high-quality datasets. In contrast, violin AMT remains underexplored due to limited annotated data. A common approach is to fine-tune pretrained models for other downstream tasks, but the effectiveness of such transfer remains unclear in the presence of timbral and articulatory differences. In this work, we investigate whether training from scratch on a medium-scale violin dataset can match the performance of fine-tuned piano-pretrained models. We adopt a piano transcription architecture without modification and train it on the MOSA dataset, which contains about 30 hours of aligned violin recordings. Our experiments on URMP and Bach10 show that models trained from scratch achieved competitive or even superior performance compared to fine-tuned counterparts. These findings suggest that strong violin AMT is possible without relying on pretrained piano representations, highlighting the importance of instrument-specific data collection and augmentation strategies.

cs.SD↗

arXiv2025

Bridging Brain Connectomes and Clinical Reports for Early Alzheimer's Disease Diagnosis

Integrating brain imaging data with clinical reports offers a valuable opportunity to leverage complementary multimodal information for more effective and timely diagnosis in practical clinical settings. This approach has gained significant attention in brain disorder research, yet a key challenge remains: how to effectively link objective imaging data with subjective text-based reports, such as doctors' notes. In this work, we propose a novel framework that aligns brain connectomes with clinical reports in a shared cross-modal latent space at both the subject and connectome levels, thereby enhancing representation learning. The key innovation of our approach is that we treat brain subnetworks as tokens of imaging data, rather than raw image patches, to align with word tokens in clinical reports. This enables a more efficient identification of system-level associations between neuroimaging findings and clinical observations, which is critical since brain disorders often manifest as network-level abnormalities rather than isolated regional alterations. We applied our method to mild cognitive impairment (MCI) using the ADNI dataset. Our approach not only achieves state-of-the-art predictive performance but also identifies clinically meaningful connectome-text pairs, offering new insights into the early mechanisms of Alzheimer's disease and supporting the development of clinically useful multimodal biomarkers.

cs.CV↗