Evaluating Theory of (an uncertain) Mind: Predicting the Uncertain Beliefs of Others in Conversation Forecasting
This addresses the challenge of quantifying uncertain beliefs in Theory of Mind for dialogue systems, though it is incremental with limited practical impact.
The paper tackles the problem of modeling the uncertainty of others' beliefs in dialogue, proposing a new suite of tasks for language models to predict this uncertainty as a probability in conversation forecasting, and finds that LMs can explain up to 7% variance in the uncertainty of others.
Typically, when evaluating Theory of Mind, we consider the beliefs of others to be binary: held or not held. But what if someone is unsure about their own beliefs? How can we quantify this uncertainty? We propose a new suite of tasks, challenging language models (LMs) to model the uncertainty of others in dialogue. We design these tasks around conversation forecasting, wherein an agent forecasts an unobserved outcome to a conversation. Uniquely, we view interlocutors themselves as forecasters, asking an LM to predict the uncertainty of the interlocutors (a probability). We experiment with re-scaling methods, variance reduction strategies, and demographic context, for this regression task, conducting experiments on three dialogue corpora (social, negotiation, task-oriented) with eight LMs. While LMs can explain up to 7% variance in the uncertainty of others, we highlight the difficulty of the tasks and room for future work, especially in practical applications, like anticipating ``false