EchoInk-R1: Exploring Audio-Visual Reasoning in Multimodal LLMs via Reinforcement Learning (🔥The Exploration of R1 for General Audio-Visual Reasoning with Qwen2.5-Omni)

OmniAgent (ICML 2026): the first native omni-modal agent for active video perception — a 7B agent that beats Qwen2.5-VL-72B with 73% fewer frames.

EchoTraffic: Enhancing Traffic Anomaly Understanding with Audio-Visual Insights (CVPR 2025)

Video Instance Shadow Detection Under the Sun and Sky (IEEE TIP 2024)

Evaluation This competition is the individual project of MSBD 5001. The individual project counts for 20% of the final grade. It's supported by the smart city project of iSingLab (Smart City). Description The dataset provides the average traffic speed per hour for a major road in Hong Kong from 2017 to 2018. Part of the dataset is provided as the training data, and your task is to predict the rest. 80% of the dataset is provided as the training data and 20% as the testing data, including the timestamp and the corresponding average speed. We sampled the testing data only from the year 2018 to provide you a training dataset that has the complete data spanning the year 2017. However, the speed information is sometimes missing due to device malfunction. You have to submit the predicted results of these testing samples, which are then compared with the ground truth to evaluate the performance of your model.