🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

A bite-size investigation into contamination we pushed out this week. Even non-score based (e.g RankNet) distillation leads to contamination so watch your GPT usage! Give the pre-print a quick read: arxiv.org/abs/2411.02284

#IR4Good paper entitled “Fair Exposure Allocation Using Generative Query Expansion” has been accepted to #ecir2025 - work by Thomas Jänich graham mcdonald and Iadh Ounis

🇮🇹🇮🇹🇮🇹 I am happy to share that our work with Sean MacAvaney and Iadh Ounis, "Improving Low-Resource Retrieval Effectiveness using Zero-Shot Linguistic Similarity Transfer", has been accepted to the #ecir2025 "IR for Good" track. Glasgow IR Group

🚨 #ECIR2025 Demo + New Package Alert 🚨 Simplifying entry into Mechanistic Interpretability in IR /w Catherine Chen Carsten Eickhoff Sean MacAvaney /🧵 soon!

A large contingent of our staff & students UofG Computing Science will soon be traveling to #ECIR2025. We'll be presenting our latest work in #search & #RecSys throughout the 5-day conference. Looking forward to connecting with colleagues and friends in Lucca! ECIR2025 UoG IDA Section, Glasgow CS

Fantastic talk from fellow TerrierTeam member Jack McKechnie in the evaluation session ECIR2025. Definitely go and check out his paper on context example selection for generated relevance assessments! #ecir2025

Another great contribution from one of my colleagues TerrierTeam 😎 Zixuan Yi opens the afternoon session ECIR2025 presenting his work on multi-modal multi-domain recommendation 🤩 Fantastic paper! 👏

Now listening to Jorge in the session on Graphs & RAG ECIR2025. Great talk 😎 #ecir2025

Awesome afternoon session here ECIR2025 so far! Next up is our very own Fangzheng Tian from Glasgow IR Group presenting a paper with the fantastic title: "Is Relevance Lost in Transmission?" Excellent work and what a venue to present it in! 🤩👏#ecir2025

Back for day 2 ECIR2025! Fantastic start with Eugene Yang presenting MURR 🤩. Great work and an excellent way to kick off the day!👏👏#ecir2025

Pasta-fuelled 🍝and back for more ECIR2025! Ariane Müller from TerrierTeam is up with Semantically Proportioned nDCG—cool work on making ColBERT more explainable 😎👏👏 #ecir2025

Next up in the IR4Good track at #ecir2025 Thomas Jänich is presenting our work (w\ Iadh Ounis) ‘Fair Exposure Allocation using Generative Query Expansion’. Glasgow IR Group UofG Computing Science

The man, the myth, the legend Thomas Jänich presenting another excellent Glasgow IR Group paper: "Fair Exposure Allocation using Generative Query Expansion" at this year's ECIR2025 #IR4GOOD track (work w/ graham mcdonald and Iadh Ounis)

Next in the IR4Good session here at #ECIR2025 is Thomas Jänich presenting his work (w/ graham mcdonald & Iadh Ounis) on fair exposure allocation. A really great talk in an interesting area! Glasgow IR Group UofG Computing Science

Our Thomas Jänich is now presenting his work entitled ‘Fair Exposure Allocation using Generative Query Expansion’ in the IR4Good session. Interesting work! #ecir2025

A morning of excellent presentations and contributions by staff and students of the Glasgow IR Group group UofG Computing Science at #ecir2025

Last day ECIR2025 but TerrierTeam isn’t done yet! Up next: our very own Andreas Chari presenting an exciting talk on Low-Resource Retrieval. Don’t miss it — and make sure to check out the paper! 😎😎👏#ecir2025

Congratulations to Thomas Jänich who today successfully defended his PhD Thesis titled ‘Fair Group-Based Exposure of Documents in Ranked Search Results’. Thanks to the examiners Sole Pera (Delft University of Technology) and Chris Anagnostopoulos! 🥳🎉👏👏 Well done Thomas! Glasgow IR Group

🎉 Glad to share that our paper "KiRAG: Knowledge-Driven Iterative Retriever for Enhancing Retrieval-Augmented Generation" (w/ Zaiqiao Meng and Craig Macdonald) has been accepted at #ACL2025 as a main conference paper!

🚨 New Pre-Print! We show that in-context learning can steer LLM rankers to satisfy multiple objectives at once, relevance and fairness/diversity, without any parameter updates. Work done jointly with Nilanjan Sinhababu from IIT Kharagpur and Debasis Ganguly. 🧵 Below!