Reading list
61 papers, by session. The title links to the publisher. The course's own copies are licensed to the class and stay on the roster, so follow the link and use your library. Syllabus
L 1. Why retrieval now Mon 31 Aug
- anchor BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models - Thakur, Reimers, Ruckle, Srivastava & Gurevych (2021), NeurIPS D&B
L 2. The inverted index Wed 02 Sep
- lecture reading Introduction to Information Retrieval - Manning, Raghavan & Schutze (2008), Cambridge University Press
L 3. Index construction and ranking foundations Mon 07 Sep
- lecture reading Introduction to Information Retrieval - Manning, Raghavan & Schutze (2008), Cambridge University Press
- lecture reading The Probabilistic Relevance Framework: BM25 and Beyond - Robertson & Zaragoza (2009), FnTIR
L 4. BM25: finish probabilistic ranking Wed 09 Sep
- lecture reading Introduction to Information Retrieval - Manning, Raghavan & Schutze (2008), Cambridge University Press
- lecture reading The Probabilistic Relevance Framework: BM25 and Beyond - Robertson & Zaragoza (2009), FnTIR
L 5. Evaluation and the changing goals of retrieval benchmarks Mon 14 Sep
- lecture reading Introduction to Information Retrieval - Manning, Raghavan & Schutze (2008), Cambridge University Press
- lecture reading Cumulated gain-based evaluation of IR techniques - Jarvelin & Kekalainen (2002), TOIS
- lecture reading Variations in relevance judgments and the measurement of retrieval effectiveness - Voorhees (2000), Information Processing & Management
L 6. Coarse retrieval and reranking Wed 16 Sep
- lecture reading Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks - Reimers & Gurevych (2019), EMNLP
- companion Dense Passage Retrieval for Open-Domain Question Answering - Karpukhin et al. (2020), EMNLP
- companion Large Dual Encoders Are Generalizable Retrievers - Ni et al. (2022), EMNLP
L 7. LLM embeddings and reranking Mon 28 Sep
- lecture reading Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models - Zhang et al. (2025), arXiv
- lecture reading Gemini Embedding: Generalizable Embeddings from Gemini - Lee et al. (2025), arXiv
- lecture reading LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders - BehnamGhader et al. (2024), COLM
- lecture reading ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT - Khattab & Zaharia (2020), SIGIR
- lecture reading Reciprocal Rank Fusion outperforms Condorcet and individual Rank Learning Methods - Cormack, Clarke & Buettcher (2009), SIGIR
D 8. BEIR Wed 30 Sep
- anchor BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models - Thakur, Reimers, Ruckle, Srivastava & Gurevych (2021), NeurIPS D&B
- companion Large Dual Encoders Are Generalizable Retrievers - Ni et al. (2022), EMNLP
- companion MS MARCO: A Human Generated MAchine Reading COmprehension Dataset - Nguyen et al. (2016), NIPS workshop
- companion Resources for Brewing BEIR: Reproducible Reference Models and Statistical Analyses - Kamalloo, Thakur, Lassance, Ma, Yang & Lin (2024), SIGIR
- companion Pyserini: A Python Toolkit for Reproducible Information Retrieval Research with Sparse and Dense Representations - Lin, Ma, Lin, Yang, Pradeep & Nogueira (2021), SIGIR
- alternate Overview of the TREC 2019 deep learning track - Craswell, Mitra, Yilmaz, Campos & Voorhees (2020), TREC
L 9. Hard queries: reasoning and ambiguity Wed 07 Oct
- lecture reading BRIGHT: A Realistic and Challenging Benchmark for Reasoning-Intensive Retrieval - Su et al. (2025), ICLR
- lecture reading ReasonIR: Training Retrievers for Reasoning Tasks - Shao et al. (2025), arXiv
- lecture reading AmbigQA: Answering Ambiguous Open-domain Questions - Min, Michael, Hajishirzi & Zettlemoyer (2020), EMNLP
- lecture reading Probabilistic Embeddings for Cross-Modal Retrieval - Chun, Oh, de Rezende, Kalantidis & Larlus (2021), CVPR
L 10. Approximate nearest neighbour search and search systems Mon 12 Oct
- lecture reading Efficient and robust approximate nearest neighbor search using Hierarchical Navigable Small World graphs - Malkov & Yashunin (2018), TPAMI
- lecture reading Product quantization for nearest neighbor search - Jegou, Douze & Schmid (2011), TPAMI
- lecture reading Billion-scale similarity search with GPUs - Johnson, Douze & Jegou (2019), IEEE Big Data
L 12. Attention as retrieval Mon 19 Oct
- anchor Retrieval Head Mechanistically Explains Long-Context Factuality - Wu, Wang, Xiao, Peng & Fu (2025), ICLR
D 13. LIMIT Wed 21 Oct
LIMIT preparation, reading sections and optional role recommendations (course sign-in).
- anchor On the Theoretical Limitations of Embedding-Based Retrieval - Weller, Boratko, Naim & Lee (2026), ICLR
- companion Is Dimensionality a Barrier for Retrieval Models? - Bangachev, Bresler, Kogan & Polyanskiy (2026), arXiv
- companion Geometrical Realization of Set Systems and Probabilistic Communication Complexity - Alon, Frankl & Rödl (1985), FOCS
- companion LIMIT: the ICLR 2026 review thread - ICLR 2026 reviewers (2026), OpenReview
- companion ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction - Santhanam, Khattab, Saad-Falcon, Potts & Zaharia (2022), NAACL
- companion Reviewer guidelines - NeurIPS (2025), conference documentation
- lecture reading Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models - Zhang et al. (2025), arXiv
- lecture reading Gemini Embedding: Generalizable Embeddings from Gemini - Lee et al. (2025), arXiv
- anchor BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models - Thakur, Reimers, Ruckle, Srivastava & Gurevych (2021), NeurIPS D&B
L 14. Memory as an architectural axis Mon 26 Oct
- lecture reading Large Memory Layers with Product Keys - Lample, Sablayrolles, Ranzato, Denoyer & Jegou (2019), NeurIPS
- anchor Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models - Cheng et al. (2026), arXiv
D 15. Retrieval Heads Wed 28 Oct
- anchor Retrieval Head Mechanistically Explains Long-Context Factuality - Wu, Wang, Xiao, Peng & Fu (2025), ICLR
- lecture reading Lost in the Middle: How Language Models Use Long Contexts - Liu et al. (2023), TACL
- lecture reading Retrieval Augmented Generation or Long-Context LLMs? A Comprehensive Study and Hybrid Approach - Li et al. (2024), EMNLP
- companion Reviewer guidelines - NeurIPS (2025), conference documentation
L 16. Memory learned at test time Mon 02 Nov
- anchor Titans: Learning to Memorize at Test Time - Behrouz, Zhong & Mirrokni (2025), NeurIPS
D 17. Engram Wed 04 Nov
- anchor Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models - Cheng et al. (2026), arXiv
- companion End-To-End Memory Networks - Sukhbaatar, Szlam, Weston & Fergus (2015), NeurIPS
- companion Memory Layers at Scale - Berges et al. (2024), arXiv
- companion Reviewer guidelines - NeurIPS (2025), conference documentation
- lecture reading Large Memory Layers with Product Keys - Lample, Sablayrolles, Ranzato, Denoyer & Jegou (2019), NeurIPS
L 18. Language models with an index attached Mon 09 Nov
- lecture reading Generalization through Memorization: Nearest Neighbor Language Models - Khandelwal, Levy, Jurafsky, Zettlemoyer & Lewis (2020), ICLR
- lecture reading Improving language models by retrieving from trillions of tokens - Borgeaud et al. (2022), ICML
- lecture reading Retrieval Augmented Generation or Long-Context LLMs? A Comprehensive Study and Hybrid Approach - Li et al. (2024), EMNLP
L 20. Retrieval-augmented generation Mon 16 Nov
- lecture reading Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks - Lewis et al. (2020), NeurIPS
- lecture reading PoisonedRAG: Knowledge Corruption Attacks to Retrieval-Augmented Generation of Large Language Models - Zou et al. (2024), USENIX Security
- lecture reading Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection - Greshake et al. (2023), AISec
D 21. Titans Wed 18 Nov
- anchor Titans: Learning to Memorize at Test Time - Behrouz, Zhong & Mirrokni (2025), NeurIPS
- companion Titans: the NeurIPS 2025 review thread - NeurIPS 2025 reviewers (2025), OpenReview
- companion Reviewer guidelines - NeurIPS (2025), conference documentation
- companion Cartridges: Lightweight and General-Purpose Long Context Representations via Self-Study - Eyuboglu et al. (2025), arXiv
- anchor Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models - Cheng et al. (2026), arXiv
L 22. Agentic retrieval and deep research Mon 23 Nov
- lecture reading ReAct: Synergizing Reasoning and Acting in Language Models - Yao et al. (2023), ICLR
- companion Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning - Jin et al. (2025), COLM
- lecture reading Evaluating Verifiability in Generative Search Engines - Liu, Zhang & Liang (2023), EMNLP Findings
L 23. Agent memory: write, manage, read Wed 25 Nov
- lecture reading Effective context engineering for AI agents - Anthropic (2025), engineering blog
- anchor Agentic Memory: Learning Unified Long-Term and Short-Term Memory Management for Large Language Model Agents - Yu, Yao, Xie, Tan, Feng, Li & Wu (2026), ACL
L 24. Learning what to remember Mon 30 Nov
- lecture reading Learning to Remember: End-to-End Training of Memory Agents for Long-Context Reasoning - Zhang et al. (2026), arXiv
- lecture reading Remember the Decision, Not the Description: A Rate-Distortion Framework for Agent Memory - Zou et al. (2026), arXiv
- lecture reading RPMem: Learning Long-Term Recurrent Parametric Memory Across Sessions for LLM Agents - Zhao et al. (2026), arXiv
- lecture reading Controlled Memory Interference in Continual LLM Agents - Ding et al. (2026), arXiv
D 25. Agentic Memory Wed 02 Dec
- anchor Agentic Memory: Learning Unified Long-Term and Short-Term Memory Management for Large Language Model Agents - Yu, Yao, Xie, Tan, Feng, Li & Wu (2026), ACL
- companion Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning - Jin et al. (2025), COLM
- companion Reviewer guidelines - NeurIPS (2025), conference documentation
- companion MemDelta: Controlled Baselines and Hidden Confounds in Agent Memory Evaluation - Wang (2026), arXiv
L 26. Evaluating agent memory Mon 07 Dec
- anchor Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions - Hu, Wang & McAuley (2026), ICLR
- lecture reading When Does Memory Help? A Cost-Aware Evaluation of Long-Term Memory in Tool-Using LLM Agents - Mishra et al. (2026), arXiv
- lecture reading DolphinBench: Mapping the Pareto Frontier of Agent Memory - Rathi et al. (2026), arXiv
- lecture reading Benchmarking AI Agent Memory: Is a Filesystem All You Need? - Letta (2025), blog
D 27. Evaluating Memory in LLM Agents Wed 09 Dec
- anchor Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions - Hu, Wang & McAuley (2026), ICLR
- companion MemDelta: Controlled Baselines and Hidden Confounds in Agent Memory Evaluation - Wang (2026), arXiv
- companion Evaluating Memory in LLM Agents: the ICLR 2026 review thread - ICLR 2026 reviewers (2026), OpenReview
- companion Reviewer guidelines - NeurIPS (2025), conference documentation
- companion Cartridges: Lightweight and General-Purpose Long Context Representations via Self-Study - Eyuboglu et al. (2025), arXiv
L 28. Privacy, personalisation and unlearning Mon 14 Dec
- lecture reading Continual Learning via Sparse Memory Finetuning - Lin et al. (2025), arXiv
- lecture reading Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection - Greshake et al. (2023), AISec