Latest events
Search events
Enter a keyword to search events within the date range above.
AIML Special Presentation: HARNESS: Hierarchical Abstractions and Reasoning for Neuro-Symbolic Robotic Systems—From Perception to Autonomy
- Date: Fri, 28 Feb 2025, 10:15 am - 11:15 am
- Location: AIML
Abstract: In this talk, I will present HARNESS: Hierarchical Abstractions and Reasoning for Neuro-Symbolic Robotic Systems—From Perception to Autonomy, showcasing our work under the DARPA Assured Neuro-symbolic Reasoning (ANSR) program. This project focuses on building fully autonomous systems that unite perception, reasoning, and planning through neuro-symbolic frameworks, emphasising explainability, performance robustness and assurance. I will first discuss NEUSIS and NEUSIS++, which were evaluated during the first and second phases of the ANSR program. These systems tackle realistic UAV search missions involving Entities of Interest (EOIs)—like “a red SUV vehicle” or “a pedestrian carrying a blue umbrella”—in complex suburban and urban settings with hazards and keep-out zones (KOZs). Their compositional neuro-symbolic design integrates visual perception, reasoning, and grounding with a probabilistic world model and a hierarchical symbolic planner. Experimental results in AirSim and Unreal Engine show they outperform a state-of-the-art (SOTA) vision-language and symbolic planning baseline in success rate, search efficiency, and 3D localisation. Next, I will introduce our recent compositional neuro-symbolic visual grounding method, NAVER, which merges advances in vision-language models, neural foundation models, probabilistic logic reasoning, and a finite-state automaton with a self-correcting mechanism. NAVER enhances robustness and interpretability and achieves SOTA performance compared to neural and neuro-symbolic baselines. If time allows, I will conclude with Hier-SLAM: Scaling-up Semantics in SLAM with a Hierarchically Categorical Gaussian Splatting, which unifies semantic SLAM under a neuro-symbolic framework—an important step toward integrated perception and mapping in the broader context of HARNESS.
AIML Research Seminar: Gaëlic Bechu
- Date: Tue, 25 Feb 2025, 10:30 am - 11:15 am
- Location: AIML
Gaëlic Bechu is a PhD student at the Centre for Augmented Reasoning, focusing on the role of AI in embodied systems, including robotics. In this seminar, he will share the latest developments in his PhD research, providing insights into his ongoing work and its applications.
AIML Special Presentation: Data-centric Computer Vision – A Practice
- Date: Mon, 24 Feb 2025, 1:30 pm - 2:30 pm
- Location: AIML
Dr Xin Yu is a senior lecturer at the University of Queensland. He is an Australian Research Council (ARC) Discovery Early Career Researcher Award 2023-2025 (DECRA) recipient and an awardee of the prestigious Google Research Scholar Program in 2021. He received his PhD degrees from both the Australian National University and Tsinghua University. His research interests involve various Computer Vision and Machine Learning topics, especially in visual data quality enhancement and recovery, human movement understanding, medical imaging analysis, and multimodal data understanding.
[Read more about AIML Special Presentation: Data-centric Computer Vision – A Practice]
Lifeblood Information Session
- Date: Fri, 21 Feb 2025, 11:00 am - 12:00 pm
- Location: AIML
Did you know that a plasma donation is needed in Australia every 18 seconds? Giving plasma is an easy, rewarding, and truly life-saving way we can give back to the community. Andrew Chadwick from Lifeblood Australia visited AIML to give a presentation to discuss how donors can get engaged with Lifeblood Australia, highlighting the eligibility requirements needed in order to donate plasma.
AIML Research Seminar - Professor Erik Dam
- Date: Tue, 11 Feb 2025
- Location: AIML
There are thousands of 3D medicals scans available with potential for understanding disease progression and phenotypes. However, without accurate and detailed annotations, machine learning methods are challenged. In particular, biomechanics models need a dense, anatomically meaningful coordinate system to simulate physiology or to do focal statistics across populations or across time. One such analysis is to understand progression of knee osteoarthritis, through statistics of shape models including bones, cartilages, and ligaments derived from thousands of knee MRI.
[Read more about AIML Research Seminar - Professor Erik Dam]
