@asash profile picture
Aleksandr V. Petrov @asash
Joined January 2009
  • Tweets 1,406
  • Following 265
  • Followers 686
201 Photos and videos
Aleksandr V. Petrov retweeted
Kamil Ciosek@MLciosek
17 Sep 2025
For anyone worried their LLM might be making stuff up, we made a budget‐friendly truth serum (semantic entropy Bayesian). See for yourself: youtube.com/watch?v=x_8ORGLD… Paper: arxiv.org/pdf/2504.03579

Hallucination Detection on a Budget: Efficient Bayesian Estimation of...

This video is about the research paper "Hallucination Detection on ...

youtube.com
7
4
1,138
Aleksandr V. Petrov retweeted
Sumit@_reachsumit
5 Sep 2025
LLMs for estimating positional bias in logged interaction data @asash et al. at Viator use LLMs to estimate position bias in logged user interaction data as an alternative to online experimentation. 📝arxiv.org/abs/2509.03696

LLMs for estimating positional bias in logged interaction data

Recommender and search systems commonly rely on Learning To Rank models trained on logged user interactions to order items by predicted relevance. However, such interaction data is often subject...

arxiv.org
1
4
504
Aleksandr V. Petrov retweeted
Craig Macdonald@craig_macdonald
29 Aug 2025
Thrilled to join @asash on the Recsperts podcast! Thanks @MarcelKurovski for having us. We had a blast discussing our #RecSys research & transformer-based sequential recommendation. Tune in on your favorite podcast platform recsperts.com/30

Recsperts - Recommender Systems Experts | #29: Transformers for Recommender Systems with Craig...

In episode 29 of Recsperts, I welcome Craig Macdonald, Professor of Information Retrieval at the University of Glasgow, and Aleksandr “Sasha” Petrov, PhD researcher and former applied scientist at ...

recsperts.com
2
5
20
805
Aleksandr V. Petrov retweeted
Glasgow IR Group@ir_glasgow
16 Jul 2025
Huge congratulations to @macavaney on receiving the prestigious ACM SIGIR Early Career Researcher Award in the research category! This well-deserved recognition highlights the excellence & impact of his work in the IR community 👏🎉#sigir2025 Cc @GlasgowCS @UofGlasgow @ACMSIGIR
1
14
70
7,982
Aleksandr V. Petrov retweeted
Alberto Carlo Maria Mancino@alberto_mancino
14 Jul 2025
2nd day at #SIGIR2025 ☀️ After the first keynote, a great talk by @asash on combining joint product quantization and dynamic pruning to accelerate top-k computation. No need to compute all scores anymore! 😎 Work with @craig_macdonald and @ntonellotto .
1
13
471
Just presented our work at #SIGIR2025. @craig_macdonald gave a great overview. If you have any questions, feel free to catch me during the coffee break!
Craig Macdonald@craig_macdonald
14 Jul 2025
.@Asash is talking about our #sigir2025 📄 applying dynamic pruning ideas in sequential recommended systems, using sub-id representations w/ myself and @ntonellotto
1
22
445
If you're at #SIGIR2025 and interested in large-scale RecSys, pop by my talk on Monday! I'll be presenting our paper (w/ @craig_macdonald and @ntonellotto): 'Efficient Recommendation with Millions of Items by Dynamic Pruning of Sub-Item Embeddings'. 🔗 arxiv.org/abs/2505.00560

Efficient Recommendation with Millions of Items by Dynamic Pruning...

A large item catalogue is a major challenge for deploying modern sequential recommender models, since it makes the memory footprint of the model large and increases inference latency. One...

arxiv.org
Glasgow IR Group@ir_glasgow
11 Jul 2025
📢 We're off to #SIGIR2025 in Padova! A large contingent of our students & staff will be at the main conference tutorials, workshops & #ICTIR2025. Let’s connect if you’re around! 🤝 Also: we're hiring in #FinTech — DM us if you're interested! #IR #Hiring #Research #ACMSIGIR
1
1
15
787
Aleksandr V. Petrov retweeted
Glasgow IR Group@ir_glasgow
10 Jul 2025
Huge congratulations to our brilliant PhD graduates: @tjaenich, @mvlacho1, @asash - It’s been a joy having you @GlasgowCS. We’re so proud of all you’ve achieved and can’t wait to see what amazing things you’ll do next. Wishing you all the best! 🎉👏 #PhDGraduation #PhDSuccess
3
8
42
1,198
Aleksandr V. Petrov retweeted
WSDM Conference@WSDMSocial
30 Jun 2025
We’re excited to announce that WSDM 2026 will take place in Boise, Idaho, from February 22 to 26, 2026! Stay tuned and visit 🔗wsdm-conference.org/2026/ for updates! #WSDM2026
5
6
2,066
Happy to share that the OARS@KDD2025 workshop accepts our work. In this work, we turn a sequential recommendation system into a semantic search using generative language models. Apparently, this works really well.
Krishna Acharya@kvachai
9 Jun 2025
1/9 Happy to share that our paper GLoSS: Generative Language Models with Semantic Search for Sequential Recommendation is accepted at the KDD OARS workshop! 🎉 Paper, code: github.com/krishnacharya/GLo… This is joint work with my wonderful collaborators @asash and Juba Ziani.
2
8
534
Thanks to everyone involved! Extremely happy that no corrections required!
Craig Macdonald@craig_macdonald
28 May 2025
Delighted that @asash passed his 🎓 PhD defense this morning, without corrections. Thanks to @pcastells and Nicolas Pugeault for their thorough examination of the thesis, and @mobilelearnfeed for convening the defense!
9
39
1,345
Aleksandr V. Petrov retweeted
Craig Macdonald@craig_macdonald
16 May 2025
I'm delighted to announce that I'm joining hashtag @Amazon Edinburgh as a part-time Visiting Scholar, working on recommendations
3
5
75
2,041
Visiting the office for the first time #Viator #Tripadvisor
31
705
The pre-print of our #SIGIR2025 paper is now available at arXiv: arxiv.org/abs/2505.00560! /w @craig_macdonald & @ntonellotto

Efficient Recommendation with Millions of Items by Dynamic Pruning...

A large item catalogue is a major challenge for deploying modern sequential recommender models, since it makes the memory footprint of the model large and increases inference latency. One...

arxiv.org
Sumit@_reachsumit
2 May 2025
Efficient Recommendation with Millions of Items by Dynamic Pruning of Sub-Item Embeddings @asash et al. introduce a dynamic pruning algorithm that efficiently finds top-K items without computing scores for the entire catalogue 📝arxiv.org/abs/2505.00560 👨🏽‍💻github.com/asash/recjpq_dp_p…
1
1
24
1,093
Aleksandr V. Petrov retweeted
Glasgow IR Group@ir_glasgow
11 Apr 2025
Arrivederci @ecir2025 #ecir2025 … and thanks for all the fish - see you all in Delft.
5
17
1,002
Aleksandr V. Petrov retweeted
Glasgow IR Group@ir_glasgow
11 Apr 2025
The last remaining members of the group at the closing of the #ir4good track and the #ecir2025 workshops. As we bid farewell to the charming and warm Lucca, it’s time to start new journeys including heading back to the sunny and green Glasgow!
1
7
32
1,552
Now listening to David Wardrope who presents our IR4Good paper (work done in Amazon; @kvachai is the lead author here). Paper link: arxiv.org/abs/2501.18117 #ECIR2025
1
1
14
437
Aleksandr V. Petrov retweeted
Glasgow IR Group@ir_glasgow
8 Apr 2025
Our own @asash has just finished presenting his paper entitled ‘CountNet: Utilising Repetition Counts in Sequential Recommendation’ in the Recommenders session of #ecir2025 - the paper is the outcome of his internship at Amazon
1
16
402
I really enjoyed presenting today! Will be happy to chat about the paper if you have any questions left.
Zixuan Yi@ZixuanYI_
8 Apr 2025
Listening to @asash presenting his paper entitled ‘CountNet: Utilising Repetition Counts in Sequential Recommendation’ in Recommenders session at @ecir2025.
1
17
553
Great news from #SIGIR2025 — our (/w @craig_macdonald and @ntonellotto) full paper “Efficient Recommendation with Millions of Items by Dynamic Pruning of Sub-Item Embeddings” has been accepted. Nice timing too: all co-authors are now at ECIR, so we get to celebrate together.
1
7
55
1,601
The paper continues our line of work on Large-Scale Sequential Recommendation. In this work we show that it is possible to find exact TOP-K highest scored items without exhaustively scoring full catalogue.
1
1
257
Load more