Kai‐Wei Chang
YOU?
Author Swipe
View article: Full-Duplex-Bench-v2: A Multi-Turn Evaluation Framework for Duplex Dialogue Systems with an Automated Examiner
Full-Duplex-Bench-v2: A Multi-Turn Evaluation Framework for Duplex Dialogue Systems with an Automated Examiner Open
While full-duplex speech agents enable natural, low-latency interaction by speaking and listening simultaneously, their consistency and task performance in multi-turn settings remain underexplored. We introduce Full-Duplex-Bench-v2 (FDB-v2…
View article: ContextNav: Towards Agentic Multimodal In-Context Learning
ContextNav: Towards Agentic Multimodal In-Context Learning Open
Recent advances demonstrate that multimodal large language models (MLLMs) exhibit strong multimodal in-context learning (ICL) capabilities, enabling them to adapt to novel vision-language tasks from a few contextual examples. However, exis…
View article: Conserved enhancer association of piRNAs and the implication in germ cell fate surveillance
Conserved enhancer association of piRNAs and the implication in germ cell fate surveillance Open
The PIWI-piRNA pathway is crucial for protecting genomic integrity from derepressed transposon invasion in developing germ cells after genome-wide epigenetic erasure. Extensive mouse studies show that pre-pachytene piRNAs contribute to est…
View article: VisRet: Visualization Improves Knowledge-Intensive Text-to-Image Retrieval
VisRet: Visualization Improves Knowledge-Intensive Text-to-Image Retrieval Open
Text-to-image retrieval (T2I retrieval) remains challenging because cross-modal embeddings often behave as bags of concepts and underrepresent structured visual relationships such as pose and viewpoint. We propose Visualize-then-Retrieve (…
View article: On The Landscape of Spoken Language Models: A Comprehensive Survey
On The Landscape of Spoken Language Models: A Comprehensive Survey Open
The field of spoken language processing is undergoing a shift from training custom-built, task-specific models toward using and optimizing spoken language models (SLMs) which act as universal speech processing systems. This trend is simila…
View article: The protective effects of liraglutide in reducing lipid droplets accumulation and myocardial fibrosis in diabetic cardiomyopathy
The protective effects of liraglutide in reducing lipid droplets accumulation and myocardial fibrosis in diabetic cardiomyopathy Open
View article: VISCO: Benchmarking Fine-Grained Critique and Correction Towards Self-Improvement in Visual Reasoning
VISCO: Benchmarking Fine-Grained Critique and Correction Towards Self-Improvement in Visual Reasoning Open
The ability of large vision-language models (LVLMs) to critique and correct their reasoning is an essential building block towards their self-improvement. However, a systematic analysis of such capabilities in LVLMs is still lacking. We pr…
View article: Dynamic-SUPERB Phase-2: A Collaboratively Expanding Benchmark for Measuring the Capabilities of Spoken Language Models with 180 Tasks
Dynamic-SUPERB Phase-2: A Collaboratively Expanding Benchmark for Measuring the Capabilities of Spoken Language Models with 180 Tasks Open
Multimodal foundation models, such as Gemini and ChatGPT, have revolutionized human-machine interactions by seamlessly integrating various forms of data. Developing a universal spoken language model that comprehends a wide range of natural…
View article: LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory
LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory Open
Recent large language model (LLM)-driven chat assistant systems have integrated memory components to track user-assistant chat histories, enabling more accurate and personalized responses. However, their long-term memory capabilities in su…
View article: Codec-SUPERB @ SLT 2024: A lightweight benchmark for neural audio codec models
Codec-SUPERB @ SLT 2024: A lightweight benchmark for neural audio codec models Open
Neural audio codec models are becoming increasingly important as they serve as tokenizers for audio, enabling efficient transmission or facilitating speech language modeling. The ideal neural audio codec should maintain content, paralingui…
View article: Con-ReCall: Detecting Pre-training Data in LLMs via Contrastive Decoding
Con-ReCall: Detecting Pre-training Data in LLMs via Contrastive Decoding Open
The training data in large language models is key to their success, but it also presents privacy and security risks, as it may contain sensitive information. Detecting pre-training data is crucial for mitigating these concerns. Existing me…
View article: Enhancing Large Vision Language Models with Self-Training on Image Comprehension
Enhancing Large Vision Language Models with Self-Training on Image Comprehension Open
Large vision language models (LVLMs) integrate large language models (LLMs) with pre-trained vision encoders, thereby activating the perception capability of the model to understand image inputs for different queries and conduct subsequent…
View article: MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?
MathVerse: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems? Open
The remarkable progress of Multi-modal Large Language Models (MLLMs) has garnered unparalleled attention, due to their superior performance in visual contexts. However, their capabilities in visual math problem-solving remain insufficientl…
View article: Roles of endogenous retroviral elements in the establishment and maintenance of imprinted gene expression
Roles of endogenous retroviral elements in the establishment and maintenance of imprinted gene expression Open
DNA methylation (DNAme) has long been recognized as a host defense mechanism, both in the restriction modification systems of prokaryotes as well as in the transcriptional silencing of repetitive elements in mammals. When DNAme was shown t…
View article: DACO: Towards Application-Driven and Comprehensive Data Analysis via Code Generation
DACO: Towards Application-Driven and Comprehensive Data Analysis via Code Generation Open
Data analysis is a crucial analytical process to generate in-depth studies and conclusive insights to comprehensively answer a given user query for tabular data. In this work, we aim to propose new resources and benchmarks to inspire futur…
View article: 6ER-028 Real-world treatment pattern and effectiveness of pirfenidone and nintedanib in patients with idiopathic pulmonary fibrosis: a multi-institutional study in Taiwan
6ER-028 Real-world treatment pattern and effectiveness of pirfenidone and nintedanib in patients with idiopathic pulmonary fibrosis: a multi-institutional study in Taiwan Open
Background and Importance Pirfenidone and nintedanib have been proven survival benefits and been currently approved for idiopathic pulmonary fibrosis (IPF). However, real-world comparison of effectiveness between two antifibrotics remains …
View article: TrustLLM: Trustworthiness in Large Language Models
TrustLLM: Trustworthiness in Large Language Models Open
Large language models (LLMs), exemplified by ChatGPT, have gained considerable attention for their excellent natural language processing capabilities. Nonetheless, these LLMs present many challenges, particularly in the realm of trustworth…
View article: Comprehensive Behavioral Analysis of Carbohydrate Sulfotransferase 9 Deficient Mice
Comprehensive Behavioral Analysis of Carbohydrate Sulfotransferase 9 Deficient Mice Open
View article: Aqueous Extract of Sparganii Rhizoma and Curcumae Rhizoma Induces Apoptosis and Inhibits Migration in Human Oral Squamous Cell Carcinoma
Aqueous Extract of Sparganii Rhizoma and Curcumae Rhizoma Induces Apoptosis and Inhibits Migration in Human Oral Squamous Cell Carcinoma Open
Sparganii Rhizoma and Curcumae Rhizoma (SRCR) are natural herbs used in traditional Chinese medicine to treat tumors and activate blood circulation. Previous studies have shown that SRCR possesses notable antitumor activity; however, the m…
View article: Improving the Adversarial Robustness of NLP Models by Information Bottleneck
Improving the Adversarial Robustness of NLP Models by Information Bottleneck Open
Existing studies have demonstrated that adversarial examples can be directly\nattributed to the presence of non-robust features, which are highly predictive,\nbut can be easily manipulated by adversaries to fool NLP models. In this study,\…
View article: Characterization of COPD in U.S. Primary Care: Data from a Real-Life COPD Registry
Characterization of COPD in U.S. Primary Care: Data from a Real-Life COPD Registry Open
Peer reviewed
View article: Variation in Demographic and Clinical Characteristics of COPD Patients Managed in U.S. Primary Care: Data from a Real-Life COPD Registry
Variation in Demographic and Clinical Characteristics of COPD Patients Managed in U.S. Primary Care: Data from a Real-Life COPD Registry Open
Peer reviewed
View article: Dynamically Expanded CNN Array for Video Coding
Dynamically Expanded CNN Array for Video Coding Open
Video coding is a critical step in all popular methods of streaming video. Marked progress has been made in video quality, compression, and computational efficiency. Recently, there has been an interest in finding ways to apply techniques …
View article: Lasagna: Multifaceted Protein-Protein Interaction Prediction Based on Siamese Residual RCNN
Lasagna: Multifaceted Protein-Protein Interaction Prediction Based on Siamese Residual RCNN Open
Sequence-based protein-protein interaction (PPI) prediction represents a fundamental computational biology problem. To address this problem, extensive research efforts have been made to extract predefined features from the sequences. Based…
View article: Correction to: Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development
Correction to: Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development Open
Following publication of the original article [1], the authors reported that one of the authors' names is spelled incorrectly.
View article: Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development
Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development Open
View article: Anomalous <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:msub><mml:mi>Z</mml:mi><mml:mn>2</mml:mn></mml:msub></mml:math> antiferromagnetic topological phase in pressurized <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:msub><mml:mi>SmB</mml:mi><mml:mn>6</mml:mn></mml:msub></mml:math>
Anomalous antiferromagnetic topological phase in pressurized Open
Antiferromagnetic materials, whose time-reversal symmetry is broken, can be classified into the Z2 topology if they respect some specific symmetry. Since the theoretical proposal, however, no materials have been found to host the antiferro…
View article: Allele-specific expression in a family quartet with autism reveals mono-to-biallelic switch and novel transcriptional processes of autism susceptibility genes
Allele-specific expression in a family quartet with autism reveals mono-to-biallelic switch and novel transcriptional processes of autism susceptibility genes Open
View article: Additional file 3: of Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development
Additional file 3: of Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development Open
Table S4. Genes and TEs targeted by stage-enriched piRNA cluster-derived piRNAs. (XLS 69 kb)
View article: Additional file 4: of Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development
Additional file 4: of Stage-dependent piRNAs in chicken implicated roles in modulating male germ cell development Open
Table S5. Number of piRNAs (in piRPM) mapped to TEs embedded in the transcripts that are highly associated with piRNAs enriched in embryonic (E11 and E14) gonadal piRNA clusters (EG-piRC). (XLS 74 kb)