arXiv subjects

Tsung-Han Wu

Publications and source records attributed to Tsung-Han Wu.

At least 19 recordsLinked to original sources

arXiv2026

CoVA-SFT: A Large-Scale Dataset for Chain of Visual Abstractions

Chain-of-thought (CoT) reasoning has dramatically improved large language models (LLMs) by allowing them to decompose problems into intermediate steps. While CoT is widely effective for linguistic tasks, text-only CoT forces models to serialize visual problems into awkward prose. Although architectural solutions exist to process visual inputs, the community lacks a massive, multi-step, self-corrected dataset to teach models how to build and maintain internal visual workspaces when solving purely textual reasoning problems. To address this limitation, we introduce CoVA-SFT, a highly structured corpus of 51.9K samples containing over 222K multimodal reasoning steps across 5 distinct layout families and 17 complex tasks, and CoVA-Bench, a companion benchmark of 1,700 held-out test samples spanning the same tasks for reproducible evaluation. By providing explicit rationale formulations, agentic renderings, and verification loops, CoVA-SFT teaches multimodal language models to interleave text and visual abstractions. We validate the dataset by demonstrating that models fine-tuned on CoVA-SFT outperform all interleaved CoT baselines by more than 2x on average on CoVA-Bench, though they still fall short of strong text-only CoT baselines, highlighting open challenges for future work.

cs.CL↗

arXiv2026

DigitalCoach: Communication and Grounding Gaps in Human and Agentic Computer Use Coaching

Agents are increasingly capable of automating software tasks, but can they teach humans how to use software themselves? We introduce DigitalCoach, a multimodal dataset of 72 human expert-novice computer use coaching sessions consisting of 22,752 dialogue turns grounded in 28.1 hours of screen and input event recordings across five software applications. We use DigitalCoach to evaluate whether state-of-the-art models can teach humans how to use computers. Automated evaluation shows that models differ from humans in how they coach: models provide more direct instructions, but fewer explanations, error diagnoses, and knowledge-check questions. When we fix the coaching method, models produce utterances similar to human references yet poorly grounded in visual context. Interactive evaluation confirms that model coaches cause learners to passively follow instructions without deeper engagement and fall short in visual grounding. DigitalCoach lays a foundation for collaborative and proactive computer use coaching agents. Data and code are available at https://project-digital-coach.vercel.app.

cs.CL↗

arXiv2026

Salami Slicing Trellis for Synchronization Errors in DNA Coding

On top of substitution errors, DNA storage channels suffer from both insertions and deletions at the same time. It is therefore important to develop error-correcting codes with efficient encoders and decoders that can combat all three types of noise. This paper introduces the salami-slicing trellis, a decision-feedback trellis that computes bitwise posterior probabilities along each strand and is coupled with polar codes across strands. The decoder alternates between advancing the trellises by one position and polar-decoding the resulting cross-strand slice, feeding the decoded bits back to the trellises for the next position. Simulations suggest that the resulting coding scheme approaches the conjectured capacity of the substitution-insertion-deletion channel.

cs.IT↗

arXiv2026

Are Large Reasoning Models Interruptible?

Real-world applications of Large Reasoning Models (LRMs) often require reasoning about changing prompts or environments. In this work, we challenge the frozen world assumption and evaluate LRM robustness under two realistic dynamic scenarios: interruptions, which test the accuracy of model responses under budget-constrained outputs, and dynamic context, which tests model adaptation to in-flight changes. Across mathematics and programming benchmarks that require long-form reasoning, static evaluations consistently overestimate robustness: even state-of-the-art LRMs, which achieve high accuracy in static settings, can fail unpredictably when interrupted or exposed to changing context, with performance dropping by up to 60% when updates are introduced late in the reasoning process. Our analysis further reveals several novel failure modes, including reasoning leakage, where models fold the reasoning into their final answer when interrupted; panic, where under time pressure models abandon reasoning entirely and return incorrect answers; and self-doubt, where performance degrades when trying to incorporate updated information. Project Page: http://dynamic-lm.github.io/

cs.CL↗

arXiv2026

Integrated ytterbium gain for visible-near-infrared photonics

Rare-earth gain media form the foundation of modern optical communications, emerging quantum hardware, and ultrafast optics. While chip-scale integration can enable fiber-like, and potentially beyond-fiber, functionality with unprecedented scalability, development in the visible and near-infrared remains in its early stages. Here, we demonstrate ytterbium-based optical gain integrated into an aluminum oxide photonic platform, achieving both single-mode lasing and optical amplification in the near-infrared regime. This platform delivers optical amplification with output powers exceeding 0.5 W, an optical-to-optical conversion efficiency above 70%, and a noise figure of 3.3 dB, approaching the quantum limit for phase-insensitive amplification. Furthermore, we achieve femtosecond pulse amplification to a record peak power of 14 kW, enabling supercontinuum generation with visible dispersive waves extending from 780 to 476 nm in conjunction with nonlinear photonic devices. This platform is compatible with heterogeneous integration into standard photonic circuits, laying the foundation for scalable visible-near-infrared photonic systems, including coherent laser arrays, mode-locked lasers, optical clocks, and microwave oscillators.

physics.optics↗

arXiv2026

SAM 3: Segment Anything with Concepts

We present Segment Anything Model (SAM) 3, a unified model that detects, segments, and tracks objects in images and videos based on concept prompts, which we define as either short noun phrases (e.g., "yellow school bus"), image exemplars, or a combination of both. Promptable Concept Segmentation (PCS) takes such prompts and returns segmentation masks and unique identities for all matching object instances. To advance PCS, we build a scalable data engine that produces a high-quality dataset with 4M unique concept labels, including hard negatives, across images and videos. Our model consists of an image-level detector and a memory-based video tracker that share a single backbone. Recognition and localization are decoupled with a presence head, which boosts detection accuracy. SAM 3 doubles the accuracy of existing systems in both image and video PCS, and improves previous SAM capabilities on visual segmentation tasks. We open source SAM 3 along with our new Segment Anything with Concepts (SA-Co) benchmark for promptable concept segmentation.

cs.CV↗

arXiv2026

Search Arena: Analyzing Search-Augmented LLMs

Search-augmented language models combine web search with Large Language Models (LLMs) to improve response groundedness and freshness. However, analyzing these systems remains challenging: existing datasets are limited in scale and narrow in scope, often constrained to static, single-turn, fact-checking questions. In this work, we introduce Search Arena, a crowd-sourced, large-scale, human-preference dataset of over 24,000 paired multi-turn user interactions with search-augmented LLMs. The dataset spans diverse intents and languages, and contains full system traces with around 12,000 human preference votes. Our analysis reveals that user preferences are influenced by the number of citations, even when the cited content does not directly support the attributed claims, uncovering a gap between perceived and actual credibility. Furthermore, user preferences vary across cited sources, revealing that community-driven platforms are generally preferred and static encyclopedic sources are not always appropriate and reliable. To assess performance across different settings, we conduct cross-arena analyses by testing search-augmented LLMs in a general-purpose chat environment and conventional LLMs in search-intensive settings. We find that web search does not degrade and may even improve performance in non-search settings; however, the quality in search settings is significantly affected if solely relying on the model's parametric knowledge. We open-sourced the dataset to support future research. Our dataset and code are available at: https://github.com/lmarena/search-arena.

cs.CL↗

arXiv2026

Visible dual-comb spectroscopy across more than 100 THz with lithium niobate nanophotonic waveguides

Broadband and high-resolution spectroscopy in the visible and ultraviolet is central to advances in multiple fields, including fundamental quantum physics, biology, atmospheric science and astronomy. Traditionally, these measurements are performed with grating or Fourier-transform spectrometers using incoherent light sources. Leveraging coherent light enables powerful frequency-comb-based techniques, but is limited by the technical complexity of efficiently generating broad spectral bandwidths from relatively narrowband and spectrally distant laser sources. Current visible dual-comb spectrometers require implicit compromises between optical bandwidth, experimental simplicity, and acquisition speed. In this work, we introduce a simple and efficient dual-comb spectrometer that converts robust Er:fiber frequency combs from the near-infrared to the ultraviolet and visible with thin-film lithium niobate (TFLN) nanophotonic waveguides. Using real-time signal processing, we retrieve coherently averaged dual-comb spectra over nearly 120 THz of simultaneous bandwidth in the visible with 100 MHz spectral resolution. With these capabilities, we measure the broadband absorption spectrum of molecular iodine (I2), demonstrating the broadest visible spectral coverage of a dual-comb spectrometer to date. Additional measurements of NO2, atomic rubidium, and atomic sodium further illustrate the achievable combination of spectroscopic bandwidth, resolution, and intrinsic frequency accuracy. Our results demonstrate the powerful integration of low-power frequency combs, nonlinear nanophotonics, and digital signal processing to enable a compact, efficient and versatile approach to high-resolution mapping of complex absorption spectra across 500 THz in the UV-visible and near-infrared spectral regions for multiple applications beyond the research lab

physics.optics↗

arXiv2025

Generate, but Verify: Reducing Hallucination in Vision-Language Models with Retrospective Resampling

Vision-Language Models (VLMs) excel at visual understanding but often suffer from visual hallucinations, where they generate descriptions of nonexistent objects, actions, or concepts, posing significant risks in safety-critical applications. Existing hallucination mitigation methods typically follow one of two paradigms: generation adjustment, which modifies decoding behavior to align text with visual inputs, and post-hoc verification, where external models assess and correct outputs. While effective, generation adjustment methods often rely on heuristics and lack correction mechanisms, while post-hoc verification is complicated, typically requiring multiple models and tending to reject outputs rather than refine them. In this work, we introduce REVERSE, a unified framework that integrates hallucination-aware training with on-the-fly self-verification. By leveraging a new hallucination-verification dataset containing over 1.3M semi-synthetic samples, along with a novel inference-time retrospective resampling technique, our approach enables VLMs to both detect hallucinations during generation and dynamically revise those hallucinations. Our evaluations show that REVERSE achieves state-of-the-art hallucination reduction, outperforming the best existing methods by up to 12% on CHAIR-MSCOCO and 34% on HaloQuest. Our dataset, model, and code are available at: https://reverse-vlm.github.io.

cs.CV↗

arXiv2025

Puzzled by Puzzles: When Vision-Language Models Can't Take a Hint

Rebus puzzles, visual riddles that encode language through imagery, spatial arrangement, and symbolic substitution, pose a unique challenge to current vision-language models (VLMs). Unlike traditional image captioning or question answering tasks, rebus solving requires multi-modal abstraction, symbolic reasoning, and a grasp of cultural, phonetic and linguistic puns. In this paper, we investigate the capacity of contemporary VLMs to interpret and solve rebus puzzles by constructing a hand-generated and annotated benchmark of diverse English-language rebus puzzles, ranging from simple pictographic substitutions to spatially-dependent cues ("head" over "heels"). We analyze how different VLMs perform, and our findings reveal that while VLMs exhibit some surprising capabilities in decoding simple visual clues, they struggle significantly with tasks requiring abstract reasoning, lateral thinking, and understanding visual metaphors.

cs.CL↗

arXiv2025

CLAIR-A: Leveraging Large Language Models to Judge Audio Captions

The Automated Audio Captioning (AAC) task asks models to generate natural language descriptions of an audio input. Evaluating these machine-generated audio captions is a complex task that requires considering diverse factors, among them, auditory scene understanding, sound-object inference, temporal coherence, and the environmental context of the scene. While current methods focus on specific aspects, they often fail to provide an overall score that aligns well with human judgment. In this work, we propose CLAIR-A, a simple and flexible method that leverages the zero-shot capabilities of large language models (LLMs) to evaluate candidate audio captions by directly asking LLMs for a semantic distance score. In our evaluations, CLAIR-A better predicts human judgements of quality compared to traditional metrics, with a 5.8% relative accuracy improvement compared to the domain-specific FENSE metric and up to 11% over the best general-purpose measure on the Clotho-Eval dataset. Moreover, CLAIR-A offers more transparency by allowing the language model to explain the reasoning behind its scores, with these explanations rated up to 30% better by human evaluators than those provided by baseline methods. CLAIR-A is made publicly available at https://github.com/DavidMChan/clair-a.

cs.CL↗

arXiv2025

LISAT: Language-Instructed Segmentation Assistant for Satellite Imagery

Segmentation models can recognize a pre-defined set of objects in images. However, models that can reason over complex user queries that implicitly refer to multiple objects of interest are still in their infancy. Recent advances in reasoning segmentation--generating segmentation masks from complex, implicit query text--demonstrate that vision-language models can operate across an open domain and produce reasonable outputs. However, our experiments show that such models struggle with complex remote-sensing imagery. In this work, we introduce LISAt, a vision-language model designed to describe complex remote-sensing scenes, answer questions about them, and segment objects of interest. We trained LISAt on a new curated geospatial reasoning-segmentation dataset, GRES, with 27,615 annotations over 9,205 images, and a multimodal pretraining dataset, PreGRES, containing over 1 million question-answer pairs. LISAt outperforms existing geospatial foundation models such as RS-GPT4V by over 10.04 % (BLEU-4) on remote-sensing description tasks, and surpasses state-of-the-art open-domain models on reasoning segmentation tasks by 143.36 % (gIoU). Our model, datasets, and code are available at https://lisat-bair.github.io/LISAt/

cs.AI↗

arXiv2025

Visual Haystacks: A Vision-Centric Needle-In-A-Haystack Benchmark

Large Multimodal Models (LMMs) have made significant strides in visual question-answering for single images. Recent advancements like long-context LMMs have allowed them to ingest larger, or even multiple, images. However, the ability to process a large number of visual tokens does not guarantee effective retrieval and reasoning for multi-image question answering (MIQA), especially in real-world applications like photo album searches or satellite imagery analysis. In this work, we first assess the limitations of current benchmarks for long-context LMMs. We address these limitations by introducing a new vision-centric, long-context benchmark, "Visual Haystacks (VHs)". We comprehensively evaluate both open-source and proprietary models on VHs, and demonstrate that these models struggle when reasoning across potentially unrelated images, perform poorly on cross-image reasoning, as well as exhibit biases based on the placement of key information within the context window. Towards a solution, we introduce MIRAGE (Multi-Image Retrieval Augmented Generation), an open-source, lightweight visual-RAG framework that processes up to 10k images on a single 40G A100 GPU -- far surpassing the 1k-image limit of contemporary models. MIRAGE demonstrates up to 13% performance improvement over existing open-source LMMs on VHs, sets a new state-of-the-art on the RetVQA multi-image QA benchmark, and achieves competitive performance on single-image QA with state-of-the-art LMMs. Our dataset, model, and code are available at: https://visual-haystacks.github.io.

cs.CV↗

arXiv2024

AED: Adaptable Error Detection for Few-shot Imitation Policy

We introduce a new task called Adaptable Error Detection (AED), which aims to identify behavior errors in few-shot imitation (FSI) policies based on visual observations in novel environments. The potential to cause serious damage to surrounding areas limits the application of FSI policies in real-world scenarios. Thus, a robust system is necessary to notify operators when FSI policies are inconsistent with the intent of demonstrations. This task introduces three challenges: (1) detecting behavior errors in novel environments, (2) identifying behavior errors that occur without revealing notable changes, and (3) lacking complete temporal information of the rollout due to the necessity of online detection. However, the existing benchmarks cannot support the development of AED because their tasks do not present all these challenges. To this end, we develop a cross-domain AED benchmark, consisting of 322 base and 153 novel environments. Additionally, we propose Pattern Observer (PrObe) to address these challenges. PrObe is equipped with a powerful pattern extractor and guided by novel learning objectives to parse discernible patterns in the policy feature representations of normal or error states. Through our comprehensive evaluation, PrObe demonstrates superior capability to detect errors arising from a wide range of FSI policies, consistently surpassing strong baselines. Moreover, we conduct detailed ablations and a pilot study on error correction to validate the effectiveness of the proposed architecture design and the practicality of the AED task, respectively. The AED project page can be found at https://aed-neurips.github.io/.

cs.RO↗

arXiv2024

Revisiting Semi-supervised Adversarial Robustness via Noise-aware Online Robust Distillation

The robust self-training (RST) framework has emerged as a prominent approach for semi-supervised adversarial training. To explore the possibility of tackling more complicated tasks with even lower labeling budgets, unlike prior approaches that rely on robust pretrained models, we present SNORD - a simple yet effective framework that introduces contemporary semi-supervised learning techniques into the realm of adversarial training. By enhancing pseudo labels and managing noisy training data more effectively, SNORD showcases impressive, state-of-the-art performance across diverse datasets and labeling budgets, all without the need for pretrained models. Compared to full adversarial supervision, SNORD achieves a 90% relative robust accuracy under epsilon = 8/255 AutoAttack, requiring less than 0.1%, 2%, and 10% labels for CIFAR-10, CIFAR-100, and TinyImageNet-200, respectively. Additional experiments confirm the efficacy of each component and demonstrate the adaptability of integrating SNORD with existing adversarial pretraining strategies to further bolster robustness.

cs.LG↗

arXiv2024

WLST: Weak Labels Guided Self-training for Weakly-supervised Domain Adaptation on 3D Object Detection

In the field of domain adaptation (DA) on 3D object detection, most of the work is dedicated to unsupervised domain adaptation (UDA). Yet, without any target annotations, the performance gap between the UDA approaches and the fully-supervised approach is still noticeable, which is impractical for real-world applications. On the other hand, weakly-supervised domain adaptation (WDA) is an underexplored yet practical task that only requires few labeling effort on the target domain. To improve the DA performance in a cost-effective way, we propose a general weak labels guided self-training framework, WLST, designed for WDA on 3D object detection. By incorporating autolabeler, which can generate 3D pseudo labels from 2D bounding boxes, into the existing self-training pipeline, our method is able to generate more robust and consistent pseudo labels that would benefit the training process on the target domain. Extensive experiments demonstrate the effectiveness, robustness, and detector-agnosticism of our WLST framework. Notably, it outperforms previous state-of-the-art methods on all evaluation tasks.

cs.CV↗

arXiv2023

See, Say, and Segment: Teaching LMMs to Overcome False Premises

Current open-source Large Multimodal Models (LMMs) excel at tasks such as open-vocabulary language grounding and segmentation but can suffer under false premises when queries imply the existence of something that is not actually present in the image. We observe that existing methods that fine-tune an LMM to segment images significantly degrade their ability to reliably determine ("see") if an object is present and to interact naturally with humans ("say"), a form of catastrophic forgetting. In this work, we propose a cascading and joint training approach for LMMs to solve this task, avoiding catastrophic forgetting of previous skills. Our resulting model can "see" by detecting whether objects are present in an image, "say" by telling the user if they are not, proposing alternative queries or correcting semantic errors in the query, and finally "segment" by outputting the mask of the desired objects if they exist. Additionally, we introduce a novel False Premise Correction benchmark dataset, an extension of existing RefCOCO(+/g) referring segmentation datasets (which we call FP-RefCOCO(+/g)). The results show that our method not only detects false premises up to 55% better than existing approaches, but under false premise conditions produces relative cIOU improvements of more than 31% over baselines, and produces natural language feedback judged helpful up to 67% of the time.

cs.CV↗

arXiv2023

Self-correcting LLM-controlled Diffusion Models

Text-to-image generation has witnessed significant progress with the advent of diffusion models. Despite the ability to generate photorealistic images, current text-to-image diffusion models still often struggle to accurately interpret and follow complex input text prompts. In contrast to existing models that aim to generate images only with their best effort, we introduce Self-correcting LLM-controlled Diffusion (SLD). SLD is a framework that generates an image from the input prompt, assesses its alignment with the prompt, and performs self-corrections on the inaccuracies in the generated image. Steered by an LLM controller, SLD turns text-to-image generation into an iterative closed-loop process, ensuring correctness in the resulting image. SLD is not only training-free but can also be seamlessly integrated with diffusion models behind API access, such as DALL-E 3, to further boost the performance of state-of-the-art diffusion models. Experimental results show that our approach can rectify a majority of incorrect generations, particularly in generative numeracy, attribute binding, and spatial relationships. Furthermore, by simply adjusting the instructions to the LLM, SLD can perform image editing tasks, bridging the gap between text-to-image generation and image editing pipelines. We will make our code available for future research and applications.

cs.CV↗