Presentation
Comprehensive Integration of Hyperdimensional Computing with Deep Learning towards Neuro-Symbolic AI
DescriptionDeep learning techniques have made great strides in machine learning, especially with image-related tasks through their ability to extract high quality features. However, many question the explainability, and as a result, the trust in deep learning models which is still ongoing research. Symbolic learning models, such as Hyperdimensional (HD) computing while more interpretable and intuitive, often struggle to discern important features. We propose NSHD, a Neuro-Symbolic model that incorporates deep learning's feature extraction and HD learning for a symbolic interpretation of features.
Event Type
Research Manuscript
TimeThursday, July 13th10:55am - 11:10am PDT
Location3004, 3rd Floor
Design
Emerging Models of Computation