I gave a talk with the workshop on how the synthesis of logic and device Mastering, Specially places like statistical relational Discovering, can allow interpretability.
Weighted model counting usually assumes that weights are only specified on literals, typically necessitating the need to introduce auxillary variables. We look at a completely new method depending on psuedo-Boolean features, resulting in a more basic definition. Empirically, we also get SOTA final results.
The Lab carries out research in synthetic intelligence, by unifying Discovering and logic, that has a modern emphasis on explainability
For anyone who is attending NeurIPS this yr, you might have an interest in trying out our papers that touch on morality, causality, and interpretability. Preprints can be found on the workshop website page.
We think about the issue of how generalized programs (strategies with loops) might be deemed appropriate in unbounded and constant domains.
A consortia project on honest techniques and goverance was recognized late previous year. Information hyperlink here.
Now we have a fresh paper accepted on Understanding best linear programming targets. We consider an “implicit“ hypothesis development solution that yields great theoretical bounds. Congrats to Gini and Alex on finding this paper recognized. Preprint here.
I gave a seminar on extending the expressiveness of probabilistic relational products with first-get options, which include common quantification about infinite domains.
Link In the final week of Oct, I gave a talk informally discussing explainability and moral responsibility in synthetic intelligence. Thanks to the organizers for the invitation.
Jonathan’s paper considers a lifted approached to weighted product integration, such as circuit development. Paulius’ paper develops a evaluate-theoretic perspective on weighted design counting and proposes a method to encode conditional weights on literals analogously to conditional probabilities, which results in substantial effectiveness advancements.
In the College of Edinburgh, he directs a exploration lab on artificial intelligence, https://vaishakbelle.com/ specialising within the unification of logic and equipment Mastering, by using a latest emphasis on explainability and ethics.
The framework is relevant to a substantial class of formalisms, like probabilistic relational versions. The paper also experiments the synthesis dilemma in that context. Preprint here.
For anyone who is attending AAAI this calendar year, you might be interested in looking at our papers that touch on fairness, abstraction and generalized sum-merchandise challenges.
I gave a chat within the dangers of synthetic intelligence and investigate priorities for the Global Advancement Society.