Page 28 - Paper Library - AiPaper

Exploring ChatGPT's Capabilities, Stability, Potential and Risks in Conducting Psychological Counseling through Simulations in School Counseling

Published:11/4/2025

Conversational AI Capabilities in Psychological CounselingApplication of ChatGPT-4 in School CounselingQuantitative Performance Evaluation of Mental Health InterventioEmotion Detection and Response Stability AnalysisDesign of Low-Intensity Mental Health Support

The study explores ChatGPT4's capabilities in simulated school counseling, showing high warmth (97.5%), empathy (94.2%), and moderate stability (ICC 0.62), highlighting the need for human oversight. Future research should involve real users and multiple model comparisons.

Enhancing Conversational Agents with Theory of Mind: Aligning Beliefs, Desires, and Intentions for Human-Like Interaction

Published:2/20/2025

Enhancement of Conversational Agents with LLMsApplication of Theory of Mind in Dialogue SystemsLLaMA ExperimentationMultimodal Interaction and Mental State Alignment

This study explores enhancing LLMdriven conversational agents with Theory of Mind (ToM) for more humanlike interaction. It demonstrates that explicit manipulation of beliefs, desires, and intentions significantly improves response consistency and quality, achieving 67% and 63%

TrackVLA++: Unleashing Reasoning and Memory Capabilities in VLA Models for Embodied Visual Tracking

Published:10/8/2025

Vision-Language-Action ModelSpatial Reasoning MechanismTarget Identification MemoryLong-Horizon Consistency ModelingAutoregressive Reasoning Model

TrackVLA is a novel VisionLanguageAction model that enhances embodied visual tracking by introducing a spatial reasoning mechanism and Target Identification Memory. It effectively addresses tracking failures under severe occlusions and achieves stateoftheart performance.

011

Directed evolution of an orthogonal transcription engine for programmable gene expression in eukaryotes

Published:12/6/2024

Programmable Gene ExpressionDirected Evolution TechniquesEukaryotic Transcription EngineT7 RNA PolymeraseSynthetic Biology Applications

This study engineered an orthogonal transcription engine by fusing T7 RNA polymerase with a capping enzyme, achieving two orders of magnitude higher transcription activity in yeast and mammalian cells, enhancing programmable gene expression for synthetic biology applications.

DIVA: A Dirichlet Process Mixtures Based Incremental Deep Clustering Algorithm via Variational Auto-Encoder

Published:5/23/2023

Incremental Deep Clustering AlgorithmDirichlet Process Mixture ModelsVariational Auto-EncoderDynamic Adaptive ClusteringOnline Variational Inference

DIVA is a nonparametric incremental deep clustering framework that combines a Dirichlet Process Mixture Model with a Variational AutoEncoder, allowing dynamic clustering adjustments without needing a predefined number of clusters, outperforming existing models on dynamic data.

Deep Plug-and-Play Clustering with Unknown Number of Clusters

Deep Clustering AlgorithmsUnknown Cluster Count AdjustmentAutomatic Clustering ModuleSplit-and-Merge FrameworkUnsupervised Classification Methods

The paper introduces a plugandplay clustering module that automatically adjusts the number of clusters without prior knowledge of K, utilizing a splitandmerge framework. Experiments demonstrate it achieves stateoftheart performance across benchmark datasets.

STYLE: Improving Domain Transferability of Asking Clarification Questions in Large Language Model Powered Conversational Agents

Published:5/20/2024

LLM-powered Conversational AgentsDomain AdaptabilityClarification Question StrategiesMultidomain Search EngineContext Understanding Capability

The paper introduces STYLE, a novel method to enhance the domain transferability of clarification question strategies in LLMpowered conversational agents. It addresses the limitations of onesizefitsall approaches and shows an average search performance improvement of 10% acr

Learning to Clarify: Multi-turn Conversations with Action-Based Contrastive Self-Training

Published:6/1/2024

Multi-Turn Conversation ModelingAction-Based Contrastive Self-TrainingDialogue Policy Optimization for Large Language ModelsAmbiguity Recognition in Human-Computer InteractionUnlabeled Dialogue Training

This paper introduces ActionBased Contrastive SelfTraining (ACT) to enhance LLMs' ability in handling ambiguity. The quasionline optimization algorithm effectively learns dialogue policies in datasparse scenarios, demonstrating superior performance over traditional finetunin

Concept -- An Evaluation Protocol on Conversational Recommender Systems with System-centric and User-centric Factors

Published:4/4/2024

Evaluation Protocol for Conversational Recommender SystemsIntegration of User Experience and System PerformanceLLM-based User SimulatorUsability Issues in Recommendation SystemsSystem-Centric and User-Centric Factors

The paper introduces the "CONCEPT" evaluation protocol that integrates systemcentric and usercentric factors in conversational recommender systems. It outlines three key characteristics and six abilities, using an LLMbased user simulator to enhance usability and user experienc

Determinants of Psychology Students’ Study Satisfaction

Published:2/25/2021

Academic Satisfaction of Psychology StudentsSelf-Efficacy and Academic PersistencePerson-Environment Fit Theory in Learning ContextsLongitudinal Survey StudyInformation Acquisition and Student Satisfaction

This study investigates the determinants of psychology students' study satisfaction using PersonEnvironment Fit Theory, emphasizing selfefficacy and information level's impact on satisfaction via persistence. Findings reveal that studyrelated characteristics significantly pred

MGCFNN: A NEURAL MULTIGRID SOLVER WITH NOVEL FOURIER NEURAL NETWORK FOR HIGH WAVENUMBER HELMHOLTZ EQUATIONS

High Wavenumber Helmholtz Equation SolvingMultigrid Neural NetworkFourier Neural NetworkNeural SolverSupervised Learning Testing and Scalability

This paper introduces MGCFNN, an advanced multigrid neural solver that utilizes a novel Fourier Neural Network (FNN) for high wavenumber Helmholtz equations, demonstrating superior performance and optimal convergence up to wavenumbers of 2000.

Enhancing Diffusion-based Restoration Models via Difficulty-Adaptive Reinforcement Learning with IQA Reward

Published:11/3/2025

Diffusion Model Image RestorationIQA-Based Reward MechanismDifficulty-Adaptive Reinforcement LearningReinforcement Learning in Image RestorationMLLM-based IQA

This study presents a difficultyadaptive reinforcement learning method that enhances diffusionbased image restoration models by integrating with an Image Quality Assessment (IQA) model, demonstrating improved fidelity and restoration performance on challenging samples.

Denoising Diffusion Probabilistic Models

Published:6/20/2020

Diffusion ModelsImage SynthesisTraining-Free Acceleration MethodsCIFAR10 DatasetProgressive Lossy Decompression

The paper presents a novel denoising diffusion probabilistic model inspired by nonequilibrium thermodynamics, achieving highquality image synthesis. By training on a weighted variational bound, it establishes a new connection with denoising score matching, attaining competitive

BioCLIP: A Vision Foundation Model for the Tree of Life

Published:12/1/2023

TreeOfLife-10M Biological Image DatasetBiological Image Classification ModelTree of Life Foundation ModelApplication of Computer Vision in BiologyBiological Information Extraction Methods

The paper introduces BioCLIP, a vision foundation model for the tree of life, leveraging the largest and most diverse biology image dataset, TreeOfLife10M. BioCLIP significantly outperforms existing models in finegrained biology classification, showcasing its strengths in diver

DoraemonGPT: Toward Understanding Dynamic Scenes with Large Language Models (Exemplified as A Video Agent)

Published:1/16/2024

LLM-based Dynamic Scene UnderstandingSpatio-Temporal Querying and Reasoning for Video AgentsMonte Carlo Tree Search PlannerSymbolic Memory RepresentationApplication of LLMs in Experiments

This paper presents DoraemonGPT, an LLMdriven system for understanding dynamic scenes, overcoming the limitations of current visual agents focused on static images. By converting videos into symbolic memory and utilizing subtask tools, it enables effective spatiotemporal reaso

Self-Chained Image-Language Model for Video Localization and Question Answering

Published:5/12/2023

Self-Recurrent Video Localization and Question AnsweringBLIP-2 Based Vision-Language ModelVideo Question AnsweringTemporal Keyframe LocalizationUnlabeled Video Localization Optimization

The SeViLA framework introduces a solution for video question answering, addressing issues from uniform frame sampling. Utilizing the BLIP2 model, it efficiently combines temporal keyframe localization and QA, significantly improving performance while reducing the need for expen

VideoAgent: Long-form Video Understanding with Large Language Model as Agent

Published:3/16/2024

Long-Form Video UnderstandingLarge Language Model as AgentMultimodal Reasoning and PlanningVision-Language ModelsEgoSchema and NExT-QA Benchmarks

The study introduces VideoAgent, an agentbased system for longform video understanding that combines a large language model with visionlanguage models. It achieves zeroshot accuracies of 54.1% and 71.3% on EgoSchema and NExTQA benchmarks, respectively, using an average of on

The Typing Cure: Experiences with Large Language Model Chatbots for Mental Health Support

Published:1/26/2024

Large Language Model for Mental Health SupportTherapeutic Value AlignmentUser Experience Research on ChatbotsEthical Design Principles for AISocial Media Mental Health Discussions

This study examines the experiences of 21 individuals using large language model chatbots for mental health support, highlighting usercreated roles, cultural limitations, and associated risks. It introduces the concept of therapeutic alignment and offers ethical design recommend

013

CRS-Que: A User-centric Evaluation Framework for Conversational Recommender Systems

Published:11/2/2023

Conversational Recommender SystemsUser-Centric Evaluation FrameworkUser Experience Evaluation MetricsMusic Exploration RecommendationMobile Phone Purchase Recommendation

This paper presents CRSQue, a usercentric evaluation framework for conversational recommender systems, built on ResQue. It integrates conversationrelated UX metrics and validates its effectiveness and reliability across different scenarios, highlighting the interaction between

Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection

Published:10/18/2023

Self-Reflective Retrieval-Augmented GenerationQuality Improvement for Large Language ModelsControllable Generation and Reflection MechanismOpen-Domain Question Answering TasksFact Verification and Citation Accuracy

The SelfRAG framework enhances large language models' quality and factual accuracy through adaptive retrieval and reflection tokens. Experiments show it significantly outperforms existing models in opendomain QA, reasoning, and fact verification tasks.

541 - 560 / 980

Papers