Iadh Ounis
iadhounis.bsky.social
Professor at University of Glasgow; Researcher in Information Retrieval and Recommender Systems
👤 Joined on 15 November 2024
281
Followers
147
Following
5
Total Posts
📜 Kullanıcı Geçmişi
- 15 Kas 2024, 05:44
🎉 Bluesky'e katıldı - İlk kullanıcı adı:
@iadhounis.bsky.social
🚫 Blokladığı Kullanıcılar
❤️ Beğendiği Gönderiler
🎄 PyTerrier Advent 24/25: Removing low-quality docs can boost search quality and cut indexing costs. Our SIGIR’24 paper QT5 trains a T5 model to filter passages at indexing time—easy to integrate, and works with dense, PISA, or SPLADE indexes too.
Show Post🎄PyTerrier Advent 23/25: You’ve done retrieval, but the results seem too homogeneous. Use a diversification reranker. Shown below is the implicit MMR diversification approach, instantiated on a BM25 or dense retrieval, but even an explicit approach like xQuAD (c.f. Rodrygo Santos) is easy to write.
Show Post
A Xmas present arrived early: Our @dagstuhl.de 24352 Manifesto on #Conversational #Agents: A Framework for #Evaluation ( #CAFE) is out! doi.org/10.4230/DagM... We define CONversational Information ACcess ( #CONIAC), propose a world model,+ a Conversational Agents Framework for Evaluation (CAFE).
Show Post🎄PyTerrier Advent 20/25: You can think of every PyTerrier transformer as a function - mapping from one dataframe type to another. This makes them easily cachable, courtesy of pyterrier_caching. We have cache object for retrievers, rerankers, or even indexing-time transformations (e.g. Doc2Query)
Show Post🎄 PyTerrier Advent 22/25: A more complex pipeline—knowledge-graph–enhanced RAG from our EMNLP 2024 paper TRACE. We build a KG over retrieved docs, then use a transformer to reason over triples for better QA. This pipeline instantiation uses a cache (see 20th advent) on LLM-based KG extraction.
Show Post