Cognitive Embeddings: Inverting Behavior to Recover Beliefs 

I address the inversion problem by recovering analysts’ cognitive (belief) embeddings from text, that is, inferring latent beliefs from observed behavior rather than merely predicting that behavior. Analysts allocate limited attention differently even when exposed to the same information environment, generating systematic forecast dispersion driven by persistent belief differences. Each analyst’s information set is represented as text embeddings, with an attention mechanism identifying which elements receive focus. Parameters are estimated using forecast errors, aligning belief embeddings with realized outcomes. The model also incorporates a memory component that allows past information to influence current beliefs, linking attention and belief formation under bounded rationality. By reconstructing beliefs before predicting outcomes, the framework inverts the forecasting process and extends to other settings characterized by shared information but heterogeneous interpretation.

• Awards: Third-Year Paper Award (Yiran Fan Memorial Fellowship)

Transformer Learning Associative Memory [Link to Paper] 

This paper links 290,000 equity analyst reports to I/B/E/S forecasts and introduces a new contextual recall measure, showing that transformer models are equivalent to standard psychological models of contextual retrieval, enabling a model that explains 43% of recall behavior, far exceeding traditional belief formation models.

• Awards: Second-Year Paper Award (Liew Fama-Miller Fellowship)

• Presentations: Machine Learning in Economics Summer Conference 2025 (MLESC25) 

Shock-Restricted Markov Switching Structural VARs [Link to Paper] 

This paper introduces a novel identification procedure using a Markov-switching VAR with external shock restrictions to examine the bidirectional relationship between uncertainty and economic activity, finding that uncertainty shocks have significant long-run adverse effects, while macroeconomic uncertainty can also stimulate industrial production, supporting the growth options theory.

• Awards: CRSP Summer Paper Award