Synthetic Data for LLM Fine-tuning with ACT-R (Interview with Alessandro Oltramari) скачать в хорошем качестве

Synthetic Data for LLM Fine-tuning with ACT-R (Interview with Alessandro Oltramari)

Alessandro Oltramari is the President of the Carnegie-Bosch Institute and a Group Leader at Bosch Research. In this video, he gives an overview of how the ACT-R cognitive architecture can be leveraged to create synthetic data to train an LLM agent with a focus on industrial use-cases. However, the discussion goes beyond that and touches on a variety of aspects in AI including cognitive models and the limits of supervised fine-tuning. CBI: carnegiebosch.cmu.edu/index.html Bosch Research: https://www.bosch.com/research/ Oltramari, A. and Lebiere, C., 2012. Using ontologies in a cognitive-grounded system: automatic action recognition in video surveillance. In Proceedings of the 7th international conference on semantic technology for intelligence, defense, and security, Fairfax. https://www.academia.edu/download/822... Binz, Marcel, and Eric Schulz. "Turning large language models into cognitive models arXiv preprint arXiv:2306.03917 (2023). https://arxiv.org/abs/2306.03917 Wu, Siyu, Alessandro Oltramari, and Frank E. Ritter. "VSM-ACT-R: Toward Using Cognitive Architecture For Manufacturing Solutions." International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation. Cham: Springer Nature Switzerland, 2024. https://link.springer.com/chapter/10.... About the channel: The Neuro Symbolic Channel provides the tutorials, courses, and research results on one of the most exciting areas in artificial intelligence and machine learning. With content originally from the AI course taught at Arizona State University, this channel brings you the latest at the intersection of symbolic methods (e.g., logic programming) and deep learning. Learn about the latest algorithms, Python packages, and progress toward larger goals such as artificial general intelligence (AGI).