Ori Plonsky
@oriplonsky.bsky.social
Assistant Professor, Technion.
Judgement and Decision Making | Behavioral science | Data science | Behavioral economics | Predicting behavior | Human learning
Judgement and Decision Making | Behavioral science | Data science | Behavioral economics | Predicting behavior | Human learning
And best of all, BEAST-GB, trained on data from some experimental contexts predicts behavior in different experimental contexts, outperforming even direct empirical generalization.
8/x
8/x
July 22, 2025 at 3:17 PM
And best of all, BEAST-GB, trained on data from some experimental contexts predicts behavior in different experimental contexts, outperforming even direct empirical generalization.
8/x
8/x
In fact, on another (3rd) large dataset where BEAST miserably fails, the ML layer saves the day – BEAST-GB outperforms >50 competing models.
7/x
7/x
July 22, 2025 at 3:17 PM
In fact, on another (3rd) large dataset where BEAST miserably fails, the ML layer saves the day – BEAST-GB outperforms >50 competing models.
7/x
7/x
What if we have more data?
BEAST-GB, trained on minimal data, outperformed deep neural nets trained on 50x more data.
Using all data, BEAST-GB hit near-ceiling prediction: closing 96% of the gap between a perfect (hypothetical) model and random guessing.
5/x
BEAST-GB, trained on minimal data, outperformed deep neural nets trained on 50x more data.
Using all data, BEAST-GB hit near-ceiling prediction: closing 96% of the gap between a perfect (hypothetical) model and random guessing.
5/x
July 22, 2025 at 3:17 PM
What if we have more data?
BEAST-GB, trained on minimal data, outperformed deep neural nets trained on 50x more data.
Using all data, BEAST-GB hit near-ceiling prediction: closing 96% of the gap between a perfect (hypothetical) model and random guessing.
5/x
BEAST-GB, trained on minimal data, outperformed deep neural nets trained on 50x more data.
Using all data, BEAST-GB hit near-ceiling prediction: closing 96% of the gap between a perfect (hypothetical) model and random guessing.
5/x
BEAST? Not Prospect Theory?
Yes, BEAST (psycnet.apa.org/record/2017-...) – a behavioral model that won a previous competition. It assumes people mentally sample outcomes and choose accordingly.
Using it to integrate theory is much better than using classical models.
4/x
Yes, BEAST (psycnet.apa.org/record/2017-...) – a behavioral model that won a previous competition. It assumes people mentally sample outcomes and choose accordingly.
Using it to integrate theory is much better than using classical models.
4/x
July 22, 2025 at 3:17 PM
BEAST? Not Prospect Theory?
Yes, BEAST (psycnet.apa.org/record/2017-...) – a behavioral model that won a previous competition. It assumes people mentally sample outcomes and choose accordingly.
Using it to integrate theory is much better than using classical models.
4/x
Yes, BEAST (psycnet.apa.org/record/2017-...) – a behavioral model that won a previous competition. It assumes people mentally sample outcomes and choose accordingly.
Using it to integrate theory is much better than using classical models.
4/x