Scholar
Pietro Lesci
Google Scholar ID: uRIcVlAAAAAJ
University of Cambridge
Citations & Impact
Citations
58
H-index
5
i10-index
2
Publications
9
Co-authors
9
list available
Publications
Resume (English only)
Academic Achievements
- - Publications:
- - 'Causal Estimation of Tokenisation Bias' accepted at ACL 2025
- - 'Causal Estimation of Memorisation Profiles' recognized as Paper of the Year by Cambridge’s Department of Computer Science and Technology, and won Best Paper Award at ACL 2024
- - 'PolyPythias: Stability and Outliers across Fifty Language Model Pre-Training Runs' (first author) accepted at ICLR 2025
- - 'Self-Training Large Language Models for Tool-Use Without Demonstrations' accepted at NAACL 2025 (Findings)
- - 'AnchorAL: Computationally Efficient Active Learning for Large and Imbalanced Datasets' accepted at NAACL 2024 (main)
- - Awards:
- - Best Paper Award at ACL 2024
- - Paper of the Year Award from Cambridge’s Department of Computer Science and Technology
- - Funding from Translated’s Imminent Research Grant
- - Other Achievements:
- - Recognized as one of the Outstanding Reviewers for EMNLP 2024
- - 'Large Language Model Memorization (L2M2)' workshop proposal accepted at ACL 2025
Research Experience
- - September 2022 to present: PhD student in Computer Science at the University of Cambridge
- - January 2022: Wordify 2.0 released
- - September 2022: Joined Amazon AWS AI Labs, working on efficient dialogue state tracking
- - June 2023: First-ever paper 'Diable: Efficient Dialogue State Tracking as Operations on Tables' accepted at ACL 2023 (Findings)
Education
- - Degree: PhD
- - University: University of Cambridge
- - Advisor: Prof Andreas Vlachos
- - Time: Since 2021
- - Major: Computer Science
Background
- - Research Interests: How training data influences a model’s behavior
- - Professional Field: Causal methods, active learning, tokenization, and pre-training
- - Background: PhD student in Computer Science at the University of Cambridge, with a background in economics and over 3 years of experience across research labs, consulting firms, and international institutions, focusing on custom model training and data science solutions.
Miscellany
- - Personal Interests: Not mentioned
