Searcharxiv⌕ Search

arXiv subjects

Kyuhong Shim

Publications and source records attributed to Kyuhong Shim.

At least 19 recordsLinked to original sources

Tap2Listen: Instant Target Switching for Streaming Audio-Visual Speaker Extraction

Streaming audio-visual target speaker extraction (AV-TSE) uses a target speaker's face to extract the speaker's speech from an input audio mixture in real time. However, existing systems fix the target speaker for an entire clip, while conversation demands frequent and immediate target switches. To address this problem, we propose Tap2Listen, a streaming AV-TSE model designed for instant response to target switches. The model separates computation by modality: the target-agnostic audio encoder processes only the mixture, while the target-specific visual cues of the selected face condition only the decoder. A switch requires only a 0.12ms re-projection and takes effect immediately. We also design component-wise losses: a prediction loss for the encoder and a supervised extraction loss for the decoder. For evaluation, we measure how quickly output quality recovers after a switch. On the standard LRS3-2mix test set, Tap2Listen improves SI-SNRi from 8.85 to 9.68 dB at a comparable parameter count. After a switch to a known face, the model produces full-quality output from the first frame, whereas the baseline needs about 1 second to recover. Finally, we show that the proposed streaming AV-TSE model runs in real time on a single CPU core.

cs.SD↗

StreamSE: Neural Audio Codec-based Streaming Speech Enhancement for On-device Inference

Recent generative speech enhancement (SE) models achieve high perceptual quality, but their non-causal processing and expensive computation make them difficult to use for streaming, on-device inference. We propose StreamSE, a 20M-parameter streaming SE model that operates in the latent space of a frozen pretrained neural audio codec. At each frame, StreamSE directly regresses a clean continuous latent from the noisy continuous latent while conditioning on its own predictions from previous frames. Enhancement therefore requires a single forward pass per frame and no iterative sampling or look-ahead. We further truncate the codec's residual vector quantizer (RVQ) to its coarse levels, halving the number of tokens per frame while improving robustness to reverberation. StreamSE is the only streaming generative model in our comparison, yet it matches or exceeds the perceptual quality of the offline generative baselines. Furthermore, the complete StreamSE pipeline runs in real time on a single smartphone CPU core. Our results show that continuous latent regression provides a practical route to generative SE under fully causal, on-device streaming constraints.

eess.AS↗

Block Recursive Transformers for Structured Parameter Sharing

Recursive Transformers depend not only on how many parameters are shared, but also on how shared parameters are organized across depth. Existing static schemes adopt two extreme reuse topologies: layer-wise repetition, where each layer is reused independently, and global cyclic reuse, where the entire stack is repeatedly applied as a single shared function. We argue that sharing granularity is an important factor in recursive parameter sharing. This study proposes Block Recursive Transformers (BRT), which partition layers into contiguous blocks and recursively reuse each block before moving to the next. This intermediate topology enables local refinement within blocks while preserving depth-wise transitions, without adding routing modules, auxiliary states, or extra training objectives. Across decoder-only models up to 1.7B parameters, with additional encoder-decoder and encoder-only studies, BRT improves validation negative log-likelihood and downstream accuracy over static recursive sharing baselines. Ablations and representation analyses further show that intermediate block granularity better balances local refinement and depth-wise differentiation. Project website: https://aim-skku.github.io/BRT/

cs.LG↗

DELTA-TTS: Adapting Autoregressive Model into Diffusion Language Model for Text-to-Speech

Autoregressive (AR) text-to-speech (TTS) models generate discrete speech tokens sequentially, which makes inference slow and can degrade robustness, since local errors propagate to later positions and can escalate into hallucination. This limitation stems from their left-to-right AR commitment: each token must be determined before future speech-token context is available. However, such ordering is not an inherent requirement for TTS, since the model receives the full input text before synthesis. In this paper, we introduce DELTA-TTS, a lightweight LoRA-based adaptation framework that converts a pretrained AR TTS model into a discrete diffusion language model (dLLM) for confidence-ordered speech-token decoding. To better capture the local structure of speech, DELTA-TTS incorporates a convolution module that injects local acoustic context, together with a 1/t-weighted training objective and a time-shifted inference schedule that together defer low-confidence positions to later steps. Trained on only 585 hours of LibriTTS, DELTA-TTS achieves a 1.75% WER on Seed-TTS test-en, outperforming its AR backbone while generating tokens 3.3 times faster. Further analysis shows that DELTA-TTS produces sharper text-speech alignment, increases overall decoding confidence, and mitigates the hallucinations observed in AR generation. Audio samples, inference code, and model weights are available at https://slpr-lab.github.io/DELTA-TTS.

eess.AS↗

KOMOS: Automatic Mean Opinion Score Estimation for Korean Speech

As speech generation and real-time spoken dialogue systems advance, neural mean opinion score (MOS) predictors are increasingly used to evaluate speech quality. However, many existing predictors are trained primarily on English read speech and generalize inconsistently across languages and speaking styles. In particular, their reliability on Korean speech has not been adequately validated. We propose KOMOS, a Korean MOS predictor for evaluating the overall quality of both read and spontaneous speech. To train and evaluate KOMOS, we collect 5,000 ratings from 29 native Korean listeners for 1,000 utterances spanning real recordings, synthesized speech, and degraded speech. KOMOS achieves an utterance-level Spearman rank correlation (SRCC) of 0.900, compared with 0.703 for UTMOS v1 and 0.767 for UTMOS v2. The inference code is publicly available at https://github.com/upskyy/komos and the trained model at https://huggingface.co/upskyy/komos.

eess.AS↗

RealSWE: A Compositional Evaluation of Coding Agents under Realistic User Requests

Coding agents are now commonly evaluated on the SWE-bench family of benchmarks, whose tasks are built from curated GitHub issues: long, structured, and information-rich. Real user requests, however, are typically far shorter and less structured. To characterize this gap, we define a six-category information taxonomy and four dimensions of linguistic style, and apply them to real user prompts from SWE-chat and problem statements from SWE-bench Verified and Pro. We find that requests carrying only a problem statement, alone or with limited additional context, account for 88% of real prompts but just 7% of benchmark problems. Furthermore, 87% of real prompts are casually written whereas 94% of benchmark problems are formal. Guided by these observations, we introduce RealSWE, 381 multi-variant task families derived from SWE-bench Verified and Pro. Variants within each family share the same underlying task and gold patch while differing only in information composition and linguistic style. Evaluating seven contemporary LLMs with RealSWE, we find that i) realistic inputs reduce resolution rates by 6.4 pp on average and can change model rankings. Controlled analysis further shows that ii) including Desired Behavior and Motivation significantly affects performance, whereas Environment Information and Reproduction Steps merely add tokens without measurable benefit; iii) linguistic style has only small, model-dependent effects. These findings provide actionable guidance for users and agents: explicitly stating the desired behavior and motivation, which most real prompts omit, substantially improves the LLM's software engineering performance.

cs.AI↗

Acoustic-to-Text KV Compression for Full-Duplex Speech Models

Full-duplex speech language models continuously accumulate acoustic key-value (KV) states, making long-running interactions memory-intensive. During listening, the model can finish processing an audio unit before the next arrives; we term the remaining interval listening-time slack. We propose acoustic-to-text KV compression, which introduces a transcription side channel to convert incoming speech into compact textual memory within this interval. When the cache exceeds a target budget during inference, older acoustic states are evicted while transcripts and recent acoustic context remain. We train the side channel with LoRA using cross-entropy on transcription segments. To preserve listening and speaking behavior, we apply knowledge distillation to the original model's token-level output distributions at native prediction positions. On ten-minute LongSpeech sessions, our MiniCPM-o 4.5 implementation reduces peak streaming KV-cache size by 64.6% compared with the same model without eviction. The proposed method also improves transcription, temporal question answering, and summarization over the baseline. Full-Duplex-Bench evaluations further show comparable pause-handling, turn-taking, and interruption performance.

cs.SD↗

Forget who you Forgot: Speaker Unlearning to Prevent Re-Identification in Zero-Shot Text-to-Speech

Recent zero-shot text-to-speech (ZS-TTS) systems can reproduce a speaker's voice with high fidelity from only a few seconds of reference speech, raising concerns over unauthorized voice cloning and impersonation. Speaker identity unlearning has recently emerged as an approach to selectively suppress this capability for speakers who opt out while preserving synthesis capability for other speakers. Although existing approaches reduce speaker similarity, preventing re-identification often faces severe degradation of speech quality. Motivated by this observation, we propose GUARD, a lightweight speaker identity unlearning framework that combines a learned speaker gate with speaker-agnostic activation steering on a frozen TTS backbone. The steering vectors are optimized using group-relative reward optimization to shift outputs from forget speakers toward population-level impostor similarity while preserving intelligibility and speech naturalness. On CosyVoice2, GUARD reduces forget-speaker similarity from 0.541 to 0.103 and re-identification accuracy in a 150-speaker gallery from 73.5% to 0.5%, while preserving retain-speaker reproduction. The results demonstrate that similarity reduction alone may not fully characterize successful speaker identity unlearning and highlight re-identification as a complementary criterion for its evaluation.

cs.SD↗

BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference

Large Reasoning Models (LRMs) achieve superior problem-solving through extended Chain-of-Thought (CoT) generation, but the resulting key-value (KV) cache grows linearly with sequence length and creates severe memory bottlenecks, often exceeding GPU capacity for long reasoning traces. Existing KV cache compression methods rely on recent queries to estimate future token importance, implicitly assuming these serve as reliable proxies for future attention patterns. We demonstrate that this assumption fails in long-horizon reasoning: certain decoding steps generate Thought Revisiting Tokens (TRT) that re-attend to distant previous context, such as task-solving plans formulated early in the trace. Through systematic analysis, we discover that queries corresponding to the TRT cluster into a small number of similarity groups in the embedding space. Based on this insight, we propose BeaconKV, a training-free KV cache compression method that maintains beacon queries, compact representatives for each global query cluster, to anticipate which KV pairs will be revisited without storing the entire query history. Across four open-source LRMs and diverse reasoning benchmarks, BeaconKV generally outperforms existing compression methods, achieving up to $5.8\times$ memory reduction while nearly preserving full cache accuracy and improving throughput by over $4.3\times$.

cs.LG↗

Revealing Multi-View Hallucination in Large Vision-Language Models

Large vision-language models (LVLMs) are increasingly being applied to multi-view image inputs captured from diverse viewpoints. Despite this growing use, current LVLMs often generate incorrect responses due to visual interference from non-target instances or viewpoints, a phenomenon we term multi-view hallucination (MVH). To systematically analyze this problem, we construct MVH-Bench, a benchmark comprising 4.8k question-answer pairs targeting two types of hallucination: cross-instance and cross-view. Empirical results show that MVH is prevalent across recent LVLMs. To address this issue, we propose Reference Shift Contrastive Decoding (RSCD), a training-free decoding technique that suppresses visual interference by generating negative logits through attention masking. Experiments on MVH-Bench with LLaVA-OneVision and Qwen2.5-VL demonstrate that RSCD improves performance by up to 25.7 and 94.8 points over existing hallucination mitigation methods, highlighting the effectiveness of RSCD.

cs.CV↗

Dependency-Aware Revocable Decoding for Efficient Diffusion Large Language Model Inference

Diffusion large language models (dLLMs) offer a promising alternative to autoregressive generation by decoding multiple tokens in parallel through iterative denoising. However, increasing decoding parallelism often degrades generation quality, as early errors can contaminate later contexts. Revocable decoding mitigates this issue by re-evaluating decoded tokens and remasking unreliable ones, but existing methods overlook that unreliable tokens may also corrupt the verification context itself. We identify this failure mode and propose Dependency-Aware Revocable Decoding (DARD), a training-free framework that separates tokens into masked, candidate, and unmasked states. DARD verifies candidate tokens using a selective context that excludes less reliable tokens and adaptively regulates their influence on subsequent decoding. Experiments across 12 textual and multimodal benchmarks on 3 open-source dLLMs show that DARD consistently improves the speed-quality Pareto frontier over recent revocable decoding methods, achieving a 2.71$\times$ speedup and a 4.35-point CIDEr score gain over Saber on Flickr30K.

cs.CL↗

Beyond Binary Instrument QA: Probing Instrument Grounding in Music Audio-Language Models

Recent music audio-language models achieve high accuracy on instrument question-answering benchmarks, but it remains unclear whether this reflects robust audio grounding or benchmark-specific shortcuts. In this paper, we introduce an OpenMIC-derived diagnostic benchmark sequence for instrument grounding in music audio-language models, extending binary instrument-presence QA to genre-prior-reduced examples, confusable instrument discrimination, longer audio context, and temporal localization. Across these settings, high binary QA accuracy often fails to predict model behavior: models can exhibit option-position bias, confusable-instrument errors, and temporal response bias. These results suggest that instrument grounding should be evaluated with multi-axis diagnostic benchmarks rather than a single aggregate accuracy.

cs.SD↗

WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Models

Recent decoder-only autoregressive text-to-speech (AR-TTS) models produce high-fidelity speech, but their memory and compute costs scale quadratically with sequence length due to full self-attention. In this paper, we propose WAND, Windowed Attention and Knowledge Distillation, a framework that adapts pretrained AR-TTS models to operate with constant computational and memory complexity. WAND separates the attention mechanism into two: persistent global attention over conditioning tokens and local sliding-window attention over generated tokens. To stabilize fine-tuning, we employ a curriculum learning strategy that progressively tightens the attention window. We further utilize knowledge distillation from a full-attention teacher to recover high-fidelity synthesis quality with high data efficiency. Evaluated on three modern AR-TTS models, WAND preserves the original quality while achieving up to 66.2% KV cache memory reduction and length-invariant, near-constant per-step latency.

cs.CL↗

Whisper-CD: Accurate Long-Form Speech Recognition using Multi-Negative Contrastive Decoding

Long-form speech recognition with large encoder-decoder models such as Whisper often exhibit hallucinations, repetition loops, and content omissions. These errors can accumulate and be further amplified when the previous segment's transcription is used as decoding context. We propose Whisper-CD, a training-free contrastive decoding framework that contrasts clean-audio logits against negative logits computed from three acoustically motivated perturbations: Gaussian noise injection, silence signal, and audio temporal shift. We aggregate these negatives via the log-sum-exp operator, building a unified multi-negative objective for token-by-token decoding. Across five English long-form benchmarks, Whisper-CD reduces WER by up to 24.3pp on CORAAL and shows 48% faster token generation throughput than beam search. Because Whisper-CD operates purely at inference time, it can be applied as a drop-in replacement to already-deployed Whisper systems without retraining.

cs.SD↗

Mask2Flow-TSE: Two-Stage Target Speaker Extraction with Masking and Flow Matching

Target speaker extraction (TSE) extracts the target speaker's voice from overlapping speech given a reference utterance. Existing masking-based approaches are lightweight and effective but suffer from an inability to synthesize missing content, leading to degraded perceptual quality. On the other hand, recent generative TSE models typically synthesize high-quality speech with diffusion, but require numerous iterative steps resulting in high computational costs and latency. We propose Mask2Flow-TSE, a two-stage framework combining the strengths of both paradigms. We introduce the deletion/insertion (D/I) proportion, an analytical tool that reveals early flow steps predominantly remove signal components rather than synthesize them. Based on this finding, we decouple deletion from insertion: a masking-based module handles the deletion-dominant early steps, while a single flow-matching step performs the remaining insertion for high-quality reconstruction. Specifically, the first stage uses lightweight convolution for the masking module, while the second stage employs a Diffusion Transformer (DiT) adapted for TSE with speaker conditioning. Unlike prior approaches that start from Gaussian noise, our method starts from the masked spectrogram, enabling high-quality reconstruction in a single inference step. Experiments show that Mask2Flow-TSE produces high-quality extractions with only 85M parameters and one-step inference, while preserving clean single-speaker inputs with minimal degradation.

cs.SD↗

TLDR: Compressing Audio Tokens for Efficient Autoregressive Text-to-Speech

Codec-based autoregressive (AR) speech language models have achieved strong text-to-speech (TTS) quality by modeling speech as sequences of discrete audio tokens with large pretrained backbones. However, this token-level formulation creates a structural efficiency bottleneck: speech-token sequences are much longer than text sequences, requiring the AR backbone to perform causal computation at every token position and maintain a KV cache that grows with the sequence length. We introduce TLDR, a patch-based autoregressive framework that accelerates codec-based AR-TTS by shifting the causal modeling from token-level speech sequences to patch-level sequences. TLDR groups consecutive codec tokens into compact latent patches using a lightweight compressor, models the resulting shorter patch sequence with a frozen pretrained AR-TTS backbone adapted by LoRA, and reconstructs fine-grained speech tokens within each patch using a speaker-conditioned extractor. With a patch size of 4, TLDR achieves a 1.8x inference speedup over the baseline AR-TTS model and reduces global KV-cache memory by up to 75%. Experimental results indicate that patch-level global causal modeling can be a practical way to reduce the inference cost of pretrained codec-based AR-TTS systems without replacing the existing modules.

cs.SD↗

GLASS: GRPO-Trained LoRA for Acoustic Style Steering in Zero-Shot Text-to-Speech

We propose GLASS, a framework for composable acoustic style control in zero-shot autoregressive text-to-speech (TTS) that learns controls from post-generation rewards rather than style labels. In zero-shot TTS, a speaker prompt often entangles speaker identity with prosodic attributes such as speaking rate and pitch, making it difficult to change style without changing the prompt itself. GLASS instead treats each acoustic attribute as a reward-defined control direction. For each control axis, GLASS freezes the TTS backbone and trains one lightweight LoRA adapter with Group Relative Policy Optimization (GRPO), using speech-token length and mean F0 as style rewards and WER as an intelligibility anchor. Because each control is represented as a LoRA weight update, independently trained adapters can be swapped, interpolated, and composed through linear LoRA arithmetic without retraining the backbone. Experiments on speaking rate and pitch control show targeted style shifts while preserving naturalness, speaker similarity, and intelligibility, and demonstrate smooth interpolation and multi-axis composition across independently trained adapters.

cs.SD↗

Feedback Adaptation for Retrieval-Augmented Generation

Retrieval-Augmented Generation (RAG) systems are typically evaluated under static assumptions, despite being frequently corrected through user or expert feedback in deployment. Existing evaluation protocols focus on overall accuracy and fail to capture how systems adapt after feedback is introduced. We introduce feedback adaptation as a problem setting for RAG systems, which asks how effectively and how quickly corrective feedback propagates to future queries. To make this behavior measurable, we propose two evaluation axes: correction lag, which captures the delay between feedback provision and behavioral change, and post-feedback performance, which measures reliability on semantically related queries after feedback. Using these metrics, we show that training-based approaches exhibit a trade-off between delayed correction and reliable adaptation. We further propose PatchRAG, a minimal inference-time instantiation that incorporates feedback without retraining, demonstrating immediate correction and strong post-feedback generalization under the proposed evaluation. Our results highlight feedback adaptation as a previously overlooked dimension of RAG system behavior in interactive settings.

cs.CL↗