Tobias Schnabel

Principal Researcher

Publications

View by:

- Lost in Transmission: When and Why LLMs Fail to Reason Globally
  
  Tobias Schnabel, Kiran Tomlinson, Adith Swaminathan, Jennifer Neville
  
  NeurIPS 2025 | December 2025
  
  NeurIPS Spotlight
  
  Preprint Github
- Symbolic Prompt Program Search: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization
  
  Tobias Schnabel, Jennifer Neville
  
  EMNLP | November 2024
  
  Publication
- On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots
  
  Christine Herlihy, Jennifer Neville, Tobias Schnabel, Adith Swaminathan
  
  2024 Uncertainty in Artificial Intelligence | July 2024
  
  Publication Project
- Prompts As Programs: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization
  
  Tobias Schnabel, Jennifer Neville
  
  April 2024
  
  Publication
- HINT: Integration Testing for AI-based Features with Humans in the Loop
  
  Quanze Chen, Tobias Schnabel, Besmira Nushi, Saleema Amershi
  
  International Conference on Intelligent User Interfaces | March 2022
  
  Publication Project
- Lightweight Compositional Embeddings for Incremental Streaming Recommendation
  
  Mengyue Hang, Tobias Schnabel, Longqi Yang, Jennifer Neville
  
  Machine Learning on Graphs, WSDM'22 | February 2022
  
  Publication
- SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  Transactions of the Association for Computational Linguistics | February 2022, Vol 10: pp. 163-177
  
  Publication
- Local Factor Models for Large-Scale Inductive Recommendation
  
  Longqi Yang, Tobias Schnabel, Paul Bennett, Susan Dumais
  
  RecSys | September 2021
  
  PDF
- Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  ACL-IJCNLP 2021 | July 2021
  
  Publication
- Debiasing Item-to-Item Recommendations With Small Annotated Datasets
  
  Tobias Schnabel, Paul Bennett
  
  RecSys | September 2020
  
  PDF Github
- “Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation
  
  Tobias Schnabel, Gonzalo Ramos, Saleema Amershi
  
  14th ACM Conference on Recommender Systems | September 2020
  
  PDF
- The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation
  
  Tobias Schnabel, Saleema Amershi, Paul Bennett, Peter Bailey, Thorsten Joachims
  
  Proceedings of the International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) | July 2020
  
  Publication
- Deep Generalized Method of Moments for Instrumental Variable Analysis
  
  Tobias Schnabel, Andrew Bennett, Nathan Kallus
  
  NeurIPS | December 2019
  
  Author's Version
- Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory
  
  Tobias Schnabel, Paul Bennett, Thorsten Joachims
  
  The 12th ACM International Conference on Web Search and Data Mining (WSDM 2019) | March 2019
  
  Publication
- Short-Term Satisfaction and Long-Term Coverage: Understanding How Users Tolerate Algorithmic Exploration
  
  Tobias Schnabel, Paul N. Bennett, Susan Dumais, Thorsten Joachims
  
  Proceedings of the 11th ACM International Conference on Web Search and Data Mining (WSDM '18) | February 2018
  
  Publication
- Unbiased Learning-to-Rank with Biased Feedback
  
  Thorsten Joachims, Adith Swaminathan, Tobias Schnabel
  
  2017 Web Search and Data Mining | February 2017
  
  Best Paper Award
  
  Publication
- Unbiased Comparative Evaluation of Ranking Functions
  
  Tobias Schnabel, Adith Swaminathan, Peter Frazier, Thorsten Joachims
  
  ACM International Conference on the Theory of Information Retrieval | September 2016
  
  Publication
- Recommendations as Treatments: Debiasing Learning and Evaluation
  
  Tobias Schnabel, Adith Swaminathan, Ashudeep Singh, Navin Chandak, Thorsten Joachims
  
  2016 International Conference on Machine Learning | June 2016
  
  Publication
- Using Shortlists to Support Decision Making and Improve Recommender System Performance
  
  Tobias Schnabel, Paul N. Bennett, Susan T. Dumais, Thorsten Joachims
  
  Proceedings of the 25th International World Wide Web Conference (WWW '16) | May 2016
  
  Publication Publication Project

- Lost in Transmission: When and Why LLMs Fail to Reason Globally
  
  Tobias Schnabel, Kiran Tomlinson, Adith Swaminathan, Jennifer Neville
  
  NeurIPS 2025 | December 2025
  
  NeurIPS Spotlight
  
  Preprint Github
- Lost in Transmission: When and Why LLMs Fail to Reason Globally
  
  Tobias Schnabel, Kiran Tomlinson, Adith Swaminathan, Jennifer Neville
  
  NeurIPS 2025 | December 2025
  
  NeurIPS Spotlight
  
  Preprint Github
- Symbolic Prompt Program Search: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization
  
  Tobias Schnabel, Jennifer Neville
  
  EMNLP | November 2024
  
  Publication
- On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots
  
  Christine Herlihy, Jennifer Neville, Tobias Schnabel, Adith Swaminathan
  
  2024 Uncertainty in Artificial Intelligence | July 2024
  
  Publication Project
- Prompts As Programs: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization
  
  Tobias Schnabel, Jennifer Neville
  
  April 2024
  
  Publication
- HINT: Integration Testing for AI-based Features with Humans in the Loop
  
  Quanze Chen, Tobias Schnabel, Besmira Nushi, Saleema Amershi
  
  International Conference on Intelligent User Interfaces | March 2022
  
  Publication Project
- SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  Transactions of the Association for Computational Linguistics | February 2022, Vol 10: pp. 163-177
  
  Publication
- Lightweight Compositional Embeddings for Incremental Streaming Recommendation
  
  Mengyue Hang, Tobias Schnabel, Longqi Yang, Jennifer Neville
  
  Machine Learning on Graphs, WSDM'22 | February 2022
  
  Publication
- Local Factor Models for Large-Scale Inductive Recommendation
  
  Longqi Yang, Tobias Schnabel, Paul Bennett, Susan Dumais
  
  RecSys | September 2021
  
  PDF
- Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  ACL-IJCNLP 2021 | July 2021
  
  Publication
- Debiasing Item-to-Item Recommendations With Small Annotated Datasets
  
  Tobias Schnabel, Paul Bennett
  
  RecSys | September 2020
  
  PDF Github
- “Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation
  
  Tobias Schnabel, Gonzalo Ramos, Saleema Amershi
  
  14th ACM Conference on Recommender Systems | September 2020
  
  PDF
- The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation
  
  Tobias Schnabel, Saleema Amershi, Paul Bennett, Peter Bailey, Thorsten Joachims
  
  Proceedings of the International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) | July 2020
  
  Publication
- Deep Generalized Method of Moments for Instrumental Variable Analysis
  
  Tobias Schnabel, Andrew Bennett, Nathan Kallus
  
  NeurIPS | December 2019
  
  Author's Version
- Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory
  
  Tobias Schnabel, Paul Bennett, Thorsten Joachims
  
  The 12th ACM International Conference on Web Search and Data Mining (WSDM 2019) | March 2019
  
  Publication
- Short-Term Satisfaction and Long-Term Coverage: Understanding How Users Tolerate Algorithmic Exploration
  
  Tobias Schnabel, Paul N. Bennett, Susan Dumais, Thorsten Joachims
  
  Proceedings of the 11th ACM International Conference on Web Search and Data Mining (WSDM '18) | February 2018
  
  Publication
- Unbiased Learning-to-Rank with Biased Feedback
  
  Thorsten Joachims, Adith Swaminathan, Tobias Schnabel
  
  2017 Web Search and Data Mining | February 2017
  
  Best Paper Award
  
  Publication
- Recommendations as Treatments: Debiasing Learning and Evaluation
  
  Tobias Schnabel, Adith Swaminathan, Ashudeep Singh, Navin Chandak, Thorsten Joachims
  
  2016 International Conference on Machine Learning | June 2016
  
  Publication
- Using Shortlists to Support Decision Making and Improve Recommender System Performance
  
  Tobias Schnabel, Paul N. Bennett, Susan T. Dumais, Thorsten Joachims
  
  Proceedings of the 25th International World Wide Web Conference (WWW '16) | May 2016
  
  Publication Publication Project
- On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots
  
  Christine Herlihy, Jennifer Neville, Tobias Schnabel, Adith Swaminathan
  
  2024 Uncertainty in Artificial Intelligence | July 2024
  
  Publication Project
- The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation
  
  Tobias Schnabel, Saleema Amershi, Paul Bennett, Peter Bailey, Thorsten Joachims
  
  Proceedings of the International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) | July 2020
  
  Publication
- Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory
  
  Tobias Schnabel, Paul Bennett, Thorsten Joachims
  
  The 12th ACM International Conference on Web Search and Data Mining (WSDM 2019) | March 2019
  
  Publication
- Unbiased Learning-to-Rank with Biased Feedback
  
  Thorsten Joachims, Adith Swaminathan, Tobias Schnabel
  
  2017 Web Search and Data Mining | February 2017
  
  Best Paper Award
  
  Publication
- Unbiased Comparative Evaluation of Ranking Functions
  
  Tobias Schnabel, Adith Swaminathan, Peter Frazier, Thorsten Joachims
  
  ACM International Conference on the Theory of Information Retrieval | September 2016
  
  Publication
- Recommendations as Treatments: Debiasing Learning and Evaluation
  
  Tobias Schnabel, Adith Swaminathan, Ashudeep Singh, Navin Chandak, Thorsten Joachims
  
  2016 International Conference on Machine Learning | June 2016
  
  Publication
- Using Shortlists to Support Decision Making and Improve Recommender System Performance
  
  Tobias Schnabel, Paul N. Bennett, Susan T. Dumais, Thorsten Joachims
  
  Proceedings of the 25th International World Wide Web Conference (WWW '16) | May 2016
  
  Publication Publication Project
- SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  Transactions of the Association for Computational Linguistics | February 2022, Vol 10: pp. 163-177
  
  Publication
- Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  ACL-IJCNLP 2021 | July 2021
  
  Publication
- HINT: Integration Testing for AI-based Features with Humans in the Loop
  
  Quanze Chen, Tobias Schnabel, Besmira Nushi, Saleema Amershi
  
  International Conference on Intelligent User Interfaces | March 2022
  
  Publication Project
- “Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation
  
  Tobias Schnabel, Gonzalo Ramos, Saleema Amershi
  
  14th ACM Conference on Recommender Systems | September 2020
  
  PDF
- The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation
  
  Tobias Schnabel, Saleema Amershi, Paul Bennett, Peter Bailey, Thorsten Joachims
  
  Proceedings of the International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) | July 2020
  
  Publication
- Using Shortlists to Support Decision Making and Improve Recommender System Performance
  
  Tobias Schnabel, Paul N. Bennett, Susan T. Dumais, Thorsten Joachims
  
  Proceedings of the 25th International World Wide Web Conference (WWW '16) | May 2016
  
  Publication Publication Project

- Lost in Transmission: When and Why LLMs Fail to Reason Globally
  
  Tobias Schnabel, Kiran Tomlinson, Adith Swaminathan, Jennifer Neville
  
  NeurIPS 2025 | December 2025
  
  NeurIPS Spotlight
  
  Preprint Github
- Symbolic Prompt Program Search: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization
  
  Tobias Schnabel, Jennifer Neville
  
  EMNLP | November 2024
  
  Publication
- On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots
  
  Christine Herlihy, Jennifer Neville, Tobias Schnabel, Adith Swaminathan
  
  2024 Uncertainty in Artificial Intelligence | July 2024
  
  Publication Project
- HINT: Integration Testing for AI-based Features with Humans in the Loop
  
  Quanze Chen, Tobias Schnabel, Besmira Nushi, Saleema Amershi
  
  International Conference on Intelligent User Interfaces | March 2022
  
  Publication Project
- Lightweight Compositional Embeddings for Incremental Streaming Recommendation
  
  Mengyue Hang, Tobias Schnabel, Longqi Yang, Jennifer Neville
  
  Machine Learning on Graphs, WSDM'22 | February 2022
  
  Publication
- Local Factor Models for Large-Scale Inductive Recommendation
  
  Longqi Yang, Tobias Schnabel, Paul Bennett, Susan Dumais
  
  RecSys | September 2021
  
  PDF
- Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  ACL-IJCNLP 2021 | July 2021
  
  Publication
- Debiasing Item-to-Item Recommendations With Small Annotated Datasets
  
  Tobias Schnabel, Paul Bennett
  
  RecSys | September 2020
  
  PDF Github
- “Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation
  
  Tobias Schnabel, Gonzalo Ramos, Saleema Amershi
  
  14th ACM Conference on Recommender Systems | September 2020
  
  PDF
- The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation
  
  Tobias Schnabel, Saleema Amershi, Paul Bennett, Peter Bailey, Thorsten Joachims
  
  Proceedings of the International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) | July 2020
  
  Publication
- Deep Generalized Method of Moments for Instrumental Variable Analysis
  
  Tobias Schnabel, Andrew Bennett, Nathan Kallus
  
  NeurIPS | December 2019
  
  Author's Version
- Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory
  
  Tobias Schnabel, Paul Bennett, Thorsten Joachims
  
  The 12th ACM International Conference on Web Search and Data Mining (WSDM 2019) | March 2019
  
  Publication
- Short-Term Satisfaction and Long-Term Coverage: Understanding How Users Tolerate Algorithmic Exploration
  
  Tobias Schnabel, Paul N. Bennett, Susan Dumais, Thorsten Joachims
  
  Proceedings of the 11th ACM International Conference on Web Search and Data Mining (WSDM '18) | February 2018
  
  Publication
- Unbiased Learning-to-Rank with Biased Feedback
  
  Thorsten Joachims, Adith Swaminathan, Tobias Schnabel
  
  2017 Web Search and Data Mining | February 2017
  
  Best Paper Award
  
  Publication
- Unbiased Comparative Evaluation of Ranking Functions
  
  Tobias Schnabel, Adith Swaminathan, Peter Frazier, Thorsten Joachims
  
  ACM International Conference on the Theory of Information Retrieval | September 2016
  
  Publication
- Recommendations as Treatments: Debiasing Learning and Evaluation
  
  Tobias Schnabel, Adith Swaminathan, Ashudeep Singh, Navin Chandak, Thorsten Joachims
  
  2016 International Conference on Machine Learning | June 2016
  
  Publication
- Using Shortlists to Support Decision Making and Improve Recommender System Performance
  
  Tobias Schnabel, Paul N. Bennett, Susan T. Dumais, Thorsten Joachims
  
  Proceedings of the 25th International World Wide Web Conference (WWW '16) | May 2016
  
  Publication Publication Project
- Prompts As Programs: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization
  
  Tobias Schnabel, Jennifer Neville
  
  April 2024
  
  Publication
- SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization
  
  Philippe Laban, Tobias Schnabel, Paul Bennett, Marti A. Hearst
  
  Transactions of the Association for Computational Linguistics | February 2022, Vol 10: pp. 163-177
  
  Publication

Tobias Schnabel

Publications

Lost in Transmission: When and Why LLMs Fail to Reason Globally

Symbolic Prompt Program Search: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization

On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots

Prompts As Programs: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization

HINT: Integration Testing for AI-based Features with Humans in the Loop

Lightweight Compositional Embeddings for Incremental Streaming Recommendation

SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization

Local Factor Models for Large-Scale Inductive Recommendation

Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text

Debiasing Item-to-Item Recommendations With Small Annotated Datasets

“Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation

The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation

Deep Generalized Method of Moments for Instrumental Variable Analysis

Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory

Short-Term Satisfaction and Long-Term Coverage: Understanding How Users Tolerate Algorithmic Exploration

Unbiased Learning-to-Rank with Biased Feedback

Unbiased Comparative Evaluation of Ranking Functions

Recommendations as Treatments: Debiasing Learning and Evaluation

Using Shortlists to Support Decision Making and Improve Recommender System Performance

Lost in Transmission: When and Why LLMs Fail to Reason Globally

Lost in Transmission: When and Why LLMs Fail to Reason Globally

Symbolic Prompt Program Search: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization

On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots

Prompts As Programs: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization

HINT: Integration Testing for AI-based Features with Humans in the Loop

SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization

Lightweight Compositional Embeddings for Incremental Streaming Recommendation

Local Factor Models for Large-Scale Inductive Recommendation

Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text

Debiasing Item-to-Item Recommendations With Small Annotated Datasets

“Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation

The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation

Deep Generalized Method of Moments for Instrumental Variable Analysis

Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory

Short-Term Satisfaction and Long-Term Coverage: Understanding How Users Tolerate Algorithmic Exploration

Unbiased Learning-to-Rank with Biased Feedback

Recommendations as Treatments: Debiasing Learning and Evaluation

Using Shortlists to Support Decision Making and Improve Recommender System Performance

On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots

The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation

Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory

Unbiased Learning-to-Rank with Biased Feedback

Unbiased Comparative Evaluation of Ranking Functions

Recommendations as Treatments: Debiasing Learning and Evaluation

Using Shortlists to Support Decision Making and Improve Recommender System Performance

SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization

Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text

HINT: Integration Testing for AI-based Features with Humans in the Loop

“Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation

The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation

Using Shortlists to Support Decision Making and Improve Recommender System Performance

Lost in Transmission: When and Why LLMs Fail to Reason Globally

Symbolic Prompt Program Search: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization

On Overcoming Miscalibrated Conversational Priors in LLM-based Chatbots

HINT: Integration Testing for AI-based Features with Humans in the Loop

Lightweight Compositional Embeddings for Incremental Streaming Recommendation

Local Factor Models for Large-Scale Inductive Recommendation

Keep It Simple: Unsupervised Simplification of Multi-Paragraph Text

Debiasing Item-to-Item Recommendations With Small Annotated Datasets

“Who doesn’t like dinosaurs?” Finding and Eliciting Richer Preferences for Recommendation

The Impact of More Transparent Interfaces on Behavior in Personalized Recommendation

Deep Generalized Method of Moments for Instrumental Variable Analysis

Shaping Feedback Data in Recommender Systems with Interventions Based on Information Foraging Theory

Short-Term Satisfaction and Long-Term Coverage: Understanding How Users Tolerate Algorithmic Exploration

Unbiased Learning-to-Rank with Biased Feedback

Unbiased Comparative Evaluation of Ranking Functions

Recommendations as Treatments: Debiasing Learning and Evaluation

Using Shortlists to Support Decision Making and Improve Recommender System Performance

Prompts As Programs: A Structure-Aware Approach to Efficient Compile-Time Prompt Optimization

SummaC: Re-visiting NLI-based Models for Inconsistency Detection in Summarization