Towards Learning Object Affordance Priors from Technical Texts
This work addresses the need for common sense knowledge in AI assistants to prevent naive and dangerous executions, though it is conceptual and incremental.
The paper tackles the problem of artificial assistants lacking common sense knowledge for safe everyday activities by proposing to extract high-confidence ability modality semantic relations (X can Y) from non-figurative texts to learn object affordance priors, such as usual modality and affordance estimates, but does not report concrete results or numbers.
Everyday activities performed by artificial assistants can potentially be executed naively and dangerously given their lack of common sense knowledge. This paper presents conceptual work towards obtaining prior knowledge on the usual modality (passive or active) of any given entity, and their affordance estimates, by extracting high-confidence ability modality semantic relations (X can Y relationship) from non-figurative texts, by analyzing co-occurrence of grammatical instances of subjects and verbs, and verbs and objects. The discussion includes an outline of the concept, potential and limitations, and possible feature and learning framework adoption.