Main Track Session Details主会技术报告专场详情
WAICA Main Track Session 1 (Best Paper Award Session)WAICA 主会技术报告专场 1（最佳论文奖专场）
Saturday, July 18, 2026 · 14:00–17:00 · Shanghai World Expo Exhibition & Convention Center — Meeting Room 82026年7月18日（星期六）· 14:00–17:00 · 世博展览馆 — 8 号会议室
| Time时间 | Title论文标题 |
|---|
| 14:00–14:15 | Branching Policy Optimization: Sandbox-Native Language Agent Reinforcement Learning |
| 14:15–14:30 | Disagreement and Beyond: Boosting Semi-supervised Medical Image Segmentation with Data Perturbation and Model Stabilization |
| 14:30–14:45 | Enhancing Vision Foundation Models via Multimodal Continual Pre-Training |
| 14:45–15:00 | Generative Visual Transmission: Rethinking Computation, Bandwidth, and Memory in Video Communication |
| 15:00–15:15 | Hierarchical Constrained Reinforcement Learning with Dynamic Boundary for Spatio-Temporal Vehicle-to-Grid Scheduling |
| 15:15–15:30 | MCPVerse: An Expansive, Real-World Benchmark for Agentic Tool Use |
| 15:30–15:45 | Reinforcement Learning with Physical Feedback for Stable 3D Molecular Generation |
| 15:45–16:00 | SGC-CLIP: Semantic Graph Guided Calibration for Class-Incremental Learning |
| 16:00–16:15 | StepPrune: Stepwise Sequential Token Selection for Multimodal Large Language Models |
| 16:15–16:30 | TIDAL: Tidal Intelligent Dynamic Allocation and Scheduling for Large-scale Federated GPU Clusters |
WAICA Main Track Session 2WAICA 主会技术报告专场 2
Sunday, July 19, 2026 · 09:00–12:00 · Shanghai World Expo Center — Medium Meeting Room 4312026年7月19日（星期日）· 09:00–12:00 · 世博中心 — 中型会议室 431
| Time时间 | Title论文标题 |
|---|
| 09:00–09:30 | Academician You He Keynote (Invited Talk)何友院士报告（特邀报告） |
| 09:30–09:45 | CAST: Mitigating Object Hallucination in Large Vision-Language Models via Caption-Guided Visual Attention Steering |
| 09:45–10:00 | FAST-LSD: Frequency-Spatial and Scale-Selective Transformer for Image-Event Line Segment Detection |
| 10:00–10:15 | Geometry-Aware Debiasing for Visual Question Answering via Mitigating Neural Collapse |
| 10:15–10:30 | MASD: Multi-Agent Structured Diagnosis for Zero-Shot Industrial Defect Inspection |
| 10:30–10:45 | SC2A-DKF: Structure-Constrained Color-Aware Dynamic Kernel Filtering for Out-of-Distribution Image Denoising |
| 10:45–11:00 | SG-Layout: Structured Scene Graph–Guided Layout Generation with LLMs |
| 11:00–11:15 | SkyVLaM: Multimodal Large Language Model for UAV Video Understanding in Remote Sensing |
| 11:15–11:30 | OphVLM-R1: Efficient Ophthalmic Reasoning via Curriculum RL |
WAICA Main Track Session 3WAICA 主会技术报告专场 3
Sunday, July 19, 2026 · 14:00–17:00 · Shanghai World Expo Center — Medium Meeting Room 4312026年7月19日（星期日）· 14:00–17:00 · 世博中心 — 中型会议室 431
| Time时间 | Title论文标题 |
|---|
| 14:00–14:15 | 3D-SSD: A 3D Networked Decoupled SSD Architecture for AI Workloads on High-Density NAND Flash |
| 14:15–14:30 | Byzantine Robust Distributed Learning with High-dimensional Efficiency |
| 14:30–14:45 | CATCH: Cross-Resolution Attention for Communication Throughput Prediction with Coupled Heterogeneous Features |
| 14:45–15:00 | CTM: Structure-Aware Incremental Placement for Large-Model GPU Clusters |
| 15:00–15:15 | EdgeSwapLLM: SLO-Aware Multi-Model Concurrent Inference for Edge LLM Serving |
| 15:15–15:30 | F2T: Fast Failover in LLM Training with Near-Zero Overhead State Management |
| 15:30–15:45 | FlowLM: Few-Step Language Modeling via Diffusion-to-Flow Adaptation |
| 15:45–16:00 | Gate Entropy Reveals Quantization Fragility: Zero-Cost Mixed-Precision Allocation for Attention-Residual LLMs |
| 16:00–16:15 | Morphling: Train Deep and Decode Shallow for Accelerated Speculative Decoding |
| 16:15–16:30 | Not All Tokens Are Equal: Inflation-Aware Routing for Agentic LLM Systems |
| 16:30–16:45 | From Stateless to Situated: Building a Psychological World for LLM-Based Agents |
WAICA Main Track Session 4WAICA 主会技术报告专场 4
Monday, July 20, 2026 · 09:00–12:00 · Shanghai World Expo Center — Medium Meeting Room 4312026年7月20日（星期一）· 09:00–12:00 · 世博中心 — 中型会议室 431
| Time时间 | Title论文标题 |
|---|
| 09:00–09:15 | A Deep Research Benchmark with Consensus-Derived Gold Rubrics for Distinguishing Financial Report Quality |
| 09:15–09:30 | A Renewable Benchmark for Failure Diagnosis in Research-Level Mathematical Reasoning |
| 09:30–09:45 | AIDABench: AI Data Analytics Benchmark |
| 09:45–10:00 | Anchoring Intent and Verifying Locally: Towards Robust Multi-Turn Search with LLMs |
| 10:00–10:15 | DebateArena: Tracking Pending Judgments and Disagreements in Multi-Agent Debate over Open Event Streams |
| 10:15–10:30 | MULTI-Bench: A Multi-Turn Interactive Benchmark for Assessing Emotional Intelligence Ability of Spoken Dialogue Models |
| 10:30–10:45 | ParaASR: Multi-Token Prediction for Fast and Long-Context LLM-Based Speech Recognition |
| 10:45–11:00 | PhysUniBench: A Multi-Modal Physics Reasoning Benchmark at Undergraduate Level |
| 11:00–11:15 | Step-Level Preference Learning for Generative Agents in Social Simulations |
WAICA Main Track Session 5WAICA 主会技术报告专场 5
Monday, July 20, 2026 · 09:00–12:00 · Shanghai World Expo Center — Medium Meeting Room 4302026年7月20日（星期一）· 09:00–12:00 · 世博中心 — 中型会议室 430
| Time时间 | Title论文标题 |
|---|
| 09:00–09:15 | Act, Sense, Act: Learning Active Perception from Large-Scale Egocentric Human Data |
| 09:15–09:30 | BlitzBough: An Efficient Privacy-Preserving Inference Framework for Decision Trees by Communication Optimization |
| 09:30–09:45 | HIRED: Hierarchical Interdisciplinary Research Team Formation via Dual-Layer Semantic Alignment |
| 09:45–10:00 | InternBootcamp: Boosting LLM Reasoning with Verifiable Task Scaling |
| 10:00–10:15 | Liability Frameworks for Agentic AI Systems |
| 10:15–10:30 | RoboFAC: A Comprehensive Framework for Robotic Failure Analysis and Correction |
| 10:30–10:45 | Source-Prior-Driven Selective Adaptation for Efficient Diffusion Model Finetuning |
| 10:45–11:00 | Statistical Calibration for LLM Watermark Detection Under Cross-Lingual Attacks |
| 11:00–11:15 | StepState: Sparse Step States for Multi-Step Reasoning |
| 11:15–11:30 | TAPER: Task-Adaptive Pairwise Energy Regularization for Graph Neural Networks |
WAICA Main Track Session 6WAICA 主会技术报告专场 6
Monday, July 20, 2026 · 09:00–12:00 · Shanghai World Expo Center — Medium Meeting Room 5152026年7月20日（星期一）· 09:00–12:00 · 世博中心 — 中型会议室 515
| Time时间 | Title论文标题 |
|---|
| 09:00–09:15 | A Physics-informed Synthetic X-ray Dataset for Miniature Robot Swarm Tracking |
| 09:15–09:30 | Beyond Accuracy: Can LLMs Provide Biologically Grounded Evidence in Sequence-Based Omics Tasks? |
| 09:30–09:45 | Cross-Reaction Transfer Learning: From Simple Hydrogenation to Complex Hydroformylation via Mechanistic Subspace Alignment |
| 09:45–10:00 | Developing and validating a GraphRAG-LLM for thematic analysis on transcripts in healthcare research |
| 10:00–10:15 | DyLAC: Latent Dynamics Clustering for Trajectory Data with Application to Mandibular Motion |
| 10:15–10:30 | MARTA: Mask-driven Anchored Reinforcement learning with Transformer-hypergraph Architecture for Interpretable Brain Disorder Diagnosis |
| 10:30–10:45 | RSPMA: Rollout-Stabilized Predictive Mesh Adaptation for Autoregressive Graph PDE Rollout |
| 10:45–11:00 | scMDDM: A generative cellular world model for transcriptome modeling toward virtual cells |
| 11:00–11:15 | SIGMA: Spectral-Implicit Generative Manifold Architecture for Spatiotemporal Turbulence Synthesis |
