Pith. sign in

REVIEW 2 major objections 4 minor 41 references

Text2Playlist: Generating Personalized Playlists from Text on Deezer

T0 review · 2 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A production system that turns natural-language requests into personalized playlists reports that 45% of its generated playlists are listened to in the following days, versus 27% for manually built playlists, which the authors take as…

desk verdict A credible industrial deployment write-up whose one headline number, 45% vs 27% listen rates, is not yet evidence of quality. read the letter →

arxiv 2501.05894 v1 pith:RECR3LZ7 submitted 2025-01-10 cs.IR cs.LG

Mathieu Delcluze , Antoine Khoury , Clémence Vast , Valerio Arnaudo , Léa Briand , Walid Bendada , Thomas Bouabça This is my paper · ORCID
classification cs.IRcs.LG
keywords text-to-playlistplaylistgenerationlargelanguagemodelscollaborativefilteringretrieval-augmentedmusicinformationretrievalbroadintentqueriesDeezer
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper reports on Text2Playlist, a feature Deezer built to turn broad, text-described music needs (like "rock from the 80s to dance all night") into personalized playlists in one step. The system first uses a large language model to pull explicit and implicit tags out of the prompt, then retrieves tracks from a tag-indexed catalog, reorders them by how close they are to the user's collaborative-filtering embedding, and finally passes the shortlist through another LLM pass that ranks for query fit, artist diversity, and overall quality. The central empirical claim is that in production, 45% of playlists generated by this tool get listened to in the following days, against 27% for manually assembled playlists. The authors also note that the most common requested moods are "Chill" and "Party", accounting for nearly half of mood requests. If the engagement gap is real, it would suggest that generative, query-driven playlist creation can serve the exploratory searches that ordinary search bars handle poorly.

What carries the argument

The pipeline has three stages. First, a large language model (the paper names Gemini Flash 1.5) converts the free-text query into a short set of tags covering genre, decade, mood, language, and similar attributes, including implicit ones the user did not state; the catalog itself already carries tags created by expert annotation and by audio-based and playlist-based models. Second, an Elasticsearch index retrieves tracks matching those tags, and a Faiss-based service scores each candidate against the user's collaborative-filtering embedding, reordering the list from most to least aligned with the user's history. Third, a second LLM pass takes the query plus descriptive text for each candidate and selects and orders tracks to maximize query fit, artist diversity, and playlist quality. The load-bearing mechanism is the combination of structured tag retrieval with embedding-based personalization, because it lets an LLM's language understanding select items from a structured catalog that is not itself free text.

What would settle it

A controlled deployment that randomizes which users see Text2Playlist, and that tracks listening after the novelty period and after matching users on playlist-creation history, would falsify the central claim if the 45% versus 27% gap vanished or reversed under those conditions.

Watch

Extended reading notes

Core claim

Text2Playlist was deployed on Deezer for premium users, first to 5% in July 2024 and then to 20% by October 2024, and the paper presents the listen-rate statistic as evidence that the tool works at scale. The system's design deliberately splits the problem: extracting structured tags from free text, leveraging the existing music catalog's tag coverage, personalizing with embeddings learned from listening data, and cleaning up the list with an LLM that sees both the original query and text descriptions of each candidate track. The paper takes the 45% versus 27% difference to indicate that users positively engage with the generated playlists, and it reports that the dominant requested moods, Chill and Party, together represent close to half of all mood tags extracted from user requests.

Load-bearing premise

The load-bearing assumption is that people who try Text2Playlist and people who manually build playlists are comparable, so the higher listen rate reflects the playlists themselves rather than novelty, placement, or differing intent.

Editorial extensions

If this is right

  • Generated playlists are listened to at a higher rate than manually created ones, so text-to-playlist can serve the exploratory side of music search more directly than a search bar.
  • Users will express broad music needs in natural language, and the dominant requests are mood-based (Chill and Party), so mood coverage deserves priority in catalog tagging.
  • The architecture supports an incremental rollout: it went from 5% to 20% of premium users without a model change, suggesting the pipeline scales with demand.
  • The engagement numbers imply that the existing search feature was not satisfying broad-intent queries, and that a one-step playlist creation flow closes that gap.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The 45% versus 27% listen-rate gap is probably not a clean measure of playlist quality: users who try a new AI feature may be predisposed to listen, and novelty or placement could inflate engagement; the paper reports no statistical controls for this comparison.
  • A randomized exposure test or a matched-cohort comparison of frequent playlist makers would tell whether the gap survives, and until then it is best read as a deployment signal rather than a causal effect.
  • Because tag extraction and refinement are separate LLM calls, the design could adapt to other content types with tag-like metadata, such as podcasts, audiobooks, or video playlists, without re-architecting the retrieval core.
  • The concentration of requests in Chill and Party suggests that future iterations could specialize the tag vocabulary and embedding space for mood, possibly making a mood-first query experience capture most user intent.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper presents Text2Playlist, a deployed Deezer tool that generates personalized playlists from free-text queries. The system first uses an LLM to extract explicit and implicit tags from the query, retrieves candidate tracks via Elasticsearch using catalog tags, personalizes through user-track collaborative-filtering embeddings, and finally re-ranks with an LLM. It was rolled out to 5% then 20% of premium users. The empirical section reports that 45% of generated playlists are listened to 'in the following days' versus 27% of manual playlists, plus popular mood tags (Chill, Party). The paper claims this indicates positive engagement and successful deployment.

Significance. If the 45%-versus-27% result were rigorously established, it would be a valuable demonstration of an LLM-based playlist generation feature deployed at scale, with interesting evidence about broad-intent music queries. The architecture is a pragmatic combination of LLM, metadata, collaborative filtering, and RAG, and the production deployment details are useful to practitioners. However, the current evidence base for the central claim is a single unadjusted observational metric, making the significance conditional on an evaluation that the paper does not currently provide.

major comments (2)
  1. [Section 4.2] The paper never defines the outcome used for the central comparison: 'proportion of playlists ... listened to in the following days.' Does 'listened to' mean at least one stream, a minimum play duration, a completed track, or a session-level criterion? The denominator (45% of how many generated playlists? 27% of which manual playlists?) and the observation window are also unspecified. Without these definitions and a confidence interval or statistical test, the raw percentages cannot support the claim of 'positive engagement.' Please report the exact metric definition, the sample sizes, the time window, and a comparison that accounts for user-level clustering.
  2. [Section 4.2] The comparison between generated and manual playlists is confounded by exposure and user selection. Text2Playlist is a new standalone feature offered to a self-selected subset of premium users during a ramp-up (5% then 20%), whereas manual playlists come from the standard workflow; users who choose to generate a playlist likely have different intent (e.g., immediate listening versus curation) and may be influenced by feature novelty or promotional placement. The causal claim 'positive engagement with the feature' requires at least a discussion of these confounds and, if possible, a matched or within-user controlled comparison. At minimum, the authors should acknowledge these limitations and soften the conclusion.
minor comments (4)
  1. [Section 1] There are several spacing errors in the introduction: 'narrowintent queries', 'afocus mindset', 'anavigational goal', and 'aninformational goal' should be corrected.
  2. [Section 5] The sentence 'as shown in 4.2, users often need to reformulate queries' is not supported by Section 4.2, which reports listen rates and mood tags; either add the reformulation evidence or remove or qualify the claim.
  3. [Section 4.2] For the observation that 'Chill' and 'Party' represent nearly half of the requested moods, please report the underlying counts or proportions and define the tag taxonomy used by the LLM.
  4. [Section 4.2] The note deferring further experiments to an ECIR Industry Talk makes the paper not fully self-contained; the key details needed to interpret Section 4.2 should appear in the paper or an appendix.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the central listen-rate claim is an external behavioral observation, and the self-citations are ordinary engineering dependencies, not load-bearing reductions.

full rationale

The paper's central empirical claim is the Section 4.2 listen-rate comparison (45% for Text2Playlist-generated playlists vs. 27% for manual playlists). This is an observed behavioral metric gathered after deployment, not a quantity derived from a fitted parameter or from an equation that already contains the conclusion. The system description in Section 3 uses internal Deezer components (tag annotations, CF embeddings, recommender modules) with several self-citations, but these are cited as existing engineering inputs, not as proof of the headline result; the headline result rests on independent production telemetry. The listen-rate comparison may suffer from confounds or under-specified definitions (e.g., what counts as 'listened to', exposure differences, self-selection), and the absence of statistical testing weakens the inference, but these are correctness and interpretability concerns, not circularity. No equation is shown to be equivalent to an input by construction, no fitted quantity is renamed as a prediction, and no uniqueness or ansatz is imported from the authors' prior work to force the conclusion. The paper is therefore self-contained with respect to its central empirical assertion, and no significant circularity is present.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

This paper is a systems report rather than a derivation, so it carries domain assumptions about the quality of each pipeline stage. The key unvalidated premises are LLM tag extraction accuracy, catalog metadata coverage, CF embedding validity for personalization, and the listen-rate metric as a satisfaction proxy. No new entities or fitted constants are introduced; the paper transfers pre-trained components into a production service.

assumptions (5)
  • domain assumption The tags extracted by the LLM (explicit and implicit) accurately reflect user intent for broad queries.
    Introduced in Section 3.1; if tag extraction fails, retrieval and refinement cannot recover relevant tracks. No evaluation of tag extraction quality is provided.
  • domain assumption Deezer's catalog tags (manual and from internal audio/text models) are sufficiently complete for the extracted query tags to retrieve a useful candidate pool.
    Section 3.1 relies on the tags existing for retrieval through Elasticsearch; coverage is asserted by earlier papers, not measured here.
  • domain assumption Cosine similarity between user and track CF embeddings is a valid signal for query-specific personalization.
    Section 3.2 uses embedding proximity as a proxy for preference; the paper provides no offline or online validation of this reranking stage in isolation.
  • domain assumption Listening to a generated playlist in the following days is a valid proxy for user satisfaction and playlist quality.
    Section 4.2 uses this metric as the main evidence of success; no definition or validation is given.
  • domain assumption The LLM refinement stage improves playlist quality according to prompt rules such as artist diversity and overall quality.
    Section 3.3 asserts it, with no ablation or quality measurement.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Text2Playlist: Generating Personalized Playlists from Text on Deezer." pith.science (2026). https://pith.science/paper/RECR3LZ7

@misc{pith2026250105894,
  author       = {Pith},
  title        = {Pith review of: Text2Playlist: Generating Personalized Playlists from Text on Deezer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RECR3LZ7}},
  note         = {Machine review of arXiv:2501.05894}
}
read the original abstract

The streaming service Deezer heavily relies on the search to help users navigate through its extensive music catalog. Nonetheless, it is primarily designed to find specific items and does not lead directly to a smooth listening experience. We present Text2Playlist, a stand-alone tool that addresses these limitations. Text2Playlist leverages generative AI, music information retrieval and recommendation systems to generate query-specific and personalized playlists, successfully deployed at scale.

Figures

Figures reproduced from arXiv: 2501.05894 by the authors.

Figure 1
Figure 1. Interface of the “Text2Playlist” tool on the website version of Deezer, creating a personalized playlist generated from an idea given by a text from a Deezer user. 2 Where to Explore and Save Music on Deezer To explore the Deezer catalog, users can either rely on recommendation features or use the search function for more targeted music content. In this paper, we will focus on this latter way to explore and save mus… view at source ↗
Figure 2
Figure 2. Overview of the Text2Playlist framework from Section 3, available for online requests in our production environment on Deezer. but we can also infer they may prefer “Focus” mood tracks to be able to work during their listening session. Thanks to a LLM, we deduce both explicit and implicit tags from the query [1,21,33]. Then we leverage tags already existing to describe the Deezer catalog. It comes from manual annota… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 34 canonical work pages

  1. [1]

    In: Proceedings of the 33rd ACM International Conference on Information and Knowledge Management

    Bai,X.,Wu,X.,Stojkovic,I.,Tsioutsiouliklis,K.:Leveraginglargelanguagemodels for improving keyphrase generation for contextual targeting. In: Proceedings of the 33rd ACM International Conference on Information and Knowledge Management. pp. 4349–4357 (2024)

    work page 2024

  2. [2]

    Bainbridge, D., Cunningham, S.J., Downie, J.S.: How people describe their music information needs: A grounded theory analysis of music queries (2003)

    work page 2003

  3. [3]

    In: Proceedings of the 17th ACM Conference on Recommender Systems

    Bendada, W., Bontempelli, T., Morlon, M., Chapus, B., Cador, T., Bouabça, T., Salha-Galvan, G.: Track Mix Generation on Music Streaming Services using Trans- formers. In: Proceedings of the 17th ACM Conference on Recommender Systems. pp. 112–115 (2023)

    work page 2023

  4. [4]

    In: Proceedings of the 14th ACM Con- ference on Recommender Systems

    Bendada, W., Salha, G., Bontempelli, T.: Carousel Personalization in Music Streaming Apps with Contextual Bandits. In: Proceedings of the 14th ACM Con- ference on Recommender Systems. pp. 420–425 (2020)

    work page 2020

  5. [5]

    Procedia Computer Science49, 136–146 (2015)

    Bokde, D., Girase, S., Mukhopadhyay, D.: Matrix Factorization Model in Collab- orative Filtering Algorithms: A Survey. Procedia Computer Science49, 136–146 (2015)

    work page 2015

  6. [6]

    In: Proceedings of the 16th ACM Conference on Recommender Systems

    Bontempelli, T., Chapus, B., Rigaud, F., Morlon, M., Lorant, M., Salha-Galvan, G.: Flow Moods: Recommending Music by Moods on Deezer. In: Proceedings of the 16th ACM Conference on Recommender Systems. pp. 452–455 (2022)

    work page 2022

  7. [7]

    In: European Conference on Information Retrieval

    Briand, L., Bontempelli, T., Bendada, W., Morlon, M., Rigaud, F., Chapus, B., Bouabça, T., Salha-Galvan, G.: Let’s get it started: Fostering the discoverability of new releases on deezer. In: European Conference on Information Retrieval. pp. 286–291. Springer (2024)

    work page 2024

  8. [8]

    In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Dis- covery and Data Mining

    Briand, L., Salha-Galvan, G., Bendada, W., Morlon, M., Tran, V.A.: A Semi- Personalized System for User Cold Start Recommendation on Music Streaming Apps. In: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Dis- covery and Data Mining. pp. 2601–2609 (2021)

    work page 2021

Show all 41 references
  1. [9]

    In: Proceedings of the 1st Workshop on NLP for Music and Audio (NLP4MusA)

    Choi, J., Khlif, A., Epure, E.: Prediction of user listening contexts for music playlists. In: Proceedings of the 1st Workshop on NLP for Music and Audio (NLP4MusA). pp. 23–27 (2020)

    2020

  2. [10]

    Deezer: https://www.deezer.com (2023)

    2023

  3. [11]

    arXiv preprint arXiv:1809.07276 (2018)

    Delbouys, R., Hennequin, R., Piccoli, F., Royo-Letelier, J., Moussallam, M.: Music mood detection based on audio and lyrics with deep neural net. arXiv preprint arXiv:1809.07276 (2018)

    2018 arXiv

  4. [12]

    Epure, E.V., Salha, G., Hennequin, R.: Multilingual music genre embeddings for ef- fectivecross-lingualmusicitemannotation.arXivpreprintarXiv:2009.07755(2020)

    2020 arXiv

  5. [13]

    Garcia-Gathright, J., St. Thomas, B., Hosey, C., Nazari, Z., Diaz, F.: Understand- ingandevaluatingusersatisfactionwithmusicdiscovery.In:The41stInternational ACM SIGIR Conference on Research & Development in Information Retrieval. pp. 55–64 (2018) 6 M. Delcluze, et al

    2018

  6. [14]

    arXiv preprint arXiv:2104.01778 (2021)

    Gong, Y., Chung, Y.A., Glass, J.: Ast: Audio spectrogram transformer. arXiv preprint arXiv:2104.01778 (2021)

    2021 arXiv

  7. [15]

    O’Reilly Media, Inc

    Gormley, C., Tong, Z.: Elasticsearch: the definitive guide: a distributed real-time search and analytics engine. " O’Reilly Media, Inc." (2015)

    2015

  8. [16]

    arXiv preprint arXiv:1809.07256 (2018)

    Hennequin, R., Royo-Letelier, J., Moussallam, M.: Audio based disambiguation of music genre tags. arXiv preprint arXiv:1809.07256 (2018)

    2018 arXiv

  9. [17]

    Thomas, B., Garcia-Gathright, J., Thom, J.: Just give me what i want: How people use and evaluate music search

    Hosey, C., Vujović, L., St. Thomas, B., Garcia-Gathright, J., Thom, J.: Just give me what i want: How people use and evaluate music search. In: Proceedings of the 2019 chi conference on human factors in computing systems. pp. 1–12 (2019)

    2019

  10. [18]

    Ibrahim, K.M., Epure, E.V., Peeters, G., Richard, G.: Should we consider the users in contextual music auto-tagging models? In: 21st International Society for Music Information Retrieval Conference (2020)

    2020

  11. [19]

    arXiv preprint arXiv:2211.07250 (2022)

    Ibrahim, K.M., Epure, E.V., Peeters, G., Richard, G.: Exploiting device and audio data to tag music with user-aware listening contexts. arXiv preprint arXiv:2211.07250 (2022)

    2022 arXiv

  12. [20]

    In: ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

    Ibrahim, K.M., Royo-Letelier, J., Epure, E.V., Peeters, G., Richard, G.: Audio- based auto-tagging with contextual tags for music. In: ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 16–20. IEEE (2020)

    2020

  13. [21]

    In: 2023 IEEE 13th International Conference on System Engineering and Technology (ICSET)

    Issa, B., Jasser, M.B., Chua, H.N., Hamzah, M.: A comparative study on embed- ding models for keyword extraction using keybert method. In: 2023 IEEE 13th International Conference on System Engineering and Technology (ICSET). pp. 40–45. IEEE (2023)

    2023

  14. [22]

    IEEE Transactions on Big Data7(3), 535–547 (2019)

    Johnson, J., Douze, M., Jégou, H.: Billion-Scale Similarity Search with GPUs. IEEE Transactions on Big Data7(3), 535–547 (2019)

    2019

  15. [23]

    Recommender Systems Handbook pp

    Koren, Y., Bell, R.: Advances in Collaborative Filtering. Recommender Systems Handbook pp. 77–118 (2015)

    2015

  16. [24]

    In: Proceedings of the 42nd International ACM SIGIR Conference on Re- search and Development in Information Retrieval

    Lamkhede, S., Das, S.: Challenges in search on streaming services: netflix case study. In: Proceedings of the 42nd International ACM SIGIR Conference on Re- search and Development in Information Retrieval. pp. 1371–1374 (2019)

    2019

  17. [25]

    Laplante, A.: Everyday life music information-seeking behaviour of young adults: an exploratory study (2008)

    2008

  18. [26]

    International Journal of Multimedia In- formation Retrieval 6, 99–113 (2017)

    Lee, S., Masoud, M., Balaji, J., Belkasim, S., Sunderraman, R., Moon, S.J.: A survey of tag-based information retrieval. International Journal of Multimedia In- formation Retrieval 6, 99–113 (2017)

    2017

  19. [27]

    Advances in Neural Information Processing Systems 33, 9459–9474 (2020)

    Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.t., Rocktäschel, T., et al.: Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems 33, 9459–9474 (2020)

    2020

  20. [28]

    In: The World Wide Web Conference

    Li, A., Thom, J., Chandar, P., Hosey, C., Thomas, B.S., Garcia-Gathright, J.: Search mindsets: Understanding focused and non-focused information seeking in music search. In: The World Wide Web Conference. pp. 2971–2977 (2019)

    2019

  21. [29]

    Physica A: Statistical Mechanics and its Applications 483, 398–411 (2017)

    Li, J., Tang, Y., Chen, J.: Leveraging tagging and rating for recommendation: Rmf meets weighted diffusion on tripartite graphs. Physica A: Statistical Mechanics and its Applications 483, 398–411 (2017)

    2017

  22. [30]

    In: Proceedings of the 46th International ACM SIGIR Conference on Re- search and Development in Information Retrieval

    Liao, L., Yang, G.H., Shah, C.: Proactive conversational agents in the post-chatgpt world. In: Proceedings of the 46th International ACM SIGIR Conference on Re- search and Development in Information Retrieval. pp. 3452–3455 (2023)

    2023

  23. [31]

    In: Proceedings of Text2Playlist 7 the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval

    Mao, K., Xiao, X., Zhu, J., Lu, B., Tang, R., He, X.: Item tagging for information retrieval: A tripartite graph neural network based approach. In: Proceedings of Text2Playlist 7 the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. ...

    2020

  24. [32]

    In: ICASSP 2024- 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

    Meseguer-Brocal, G., Desblancs, D., Hennequin, R.: An experimental compari- son of multi-view self-supervised methods for music tagging. In: ICASSP 2024- 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). pp. 1141–1145. IEEE (2024)

    2024

  25. [33]

    arXiv preprint arXiv:2307.06435 (2023)

    Naveed, H., Khan, A.U., Qiu, S., Saqib, M., Anwar, S., Usman, M., Akhtar, N., Barnes, N., Mian, A.: A comprehensive overview of large language models. arXiv preprint arXiv:2307.06435 (2023)

    2023 arXiv

  26. [34]

    In: Proceedings of the ACM Web Conference 2023

    Penha, G., Palumbo, E., Aziz, M., Wang, A., Bouchard, H.: Improving content retrievability in search with controllable query generation. In: Proceedings of the ACM Web Conference 2023. pp. 3182–3192 (2023)

    2023

  27. [35]

    Petroni, F., Rocktäschel, T., Lewis, P., Bakhtin, A., Wu, Y., Miller, A.H., Riedel, S.: Language models as knowledge bases? arXiv preprint arXiv:1909.01066 (2019)

    2019 arXiv

  28. [36]

    In: Proceedings of the 30th ACM Conference on User Modeling, Adaptation and Per- sonalization

    Sguerra, B., Baranes, M., Hennequin, R., Moussallam, M.: Navigational, informa- tional or punk-rock? an exploration of search intent in the musical domain. In: Proceedings of the 30th ACM Conference on User Modeling, Adaptation and Per- sonalization. pp. 202–211 (2022)

    2022

  29. [37]

    In: Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining

    Su, N., He, J., Liu, Y., Zhang, M., Ma, S.: User intent, behaviour, and perceived satisfaction in product search. In: Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining. pp. 547–555 (2018)

    2018

  30. [38]

    arXiv preprint arXiv:2403.05530 (2024)

    Team, G., Georgiev, P., Lei, V.I., Burnell, R., Bai, L., Gulati, A., Tanzer, G., Vin- cent,D.,Pan,Z.,Wang,S.,etal.:Gemini1.5:Unlockingmultimodalunderstanding across millions of tokens of context. arXiv preprint arXiv:2403.05530 (2024)

    2024 arXiv

  31. [39]

    In: Proceedings of the 13th ACM conference on recommender systems

    Yi, X., Yang, J., Hong, L., Cheng, D.Z., Heldt, L., Kumthekar, A., Zhao, Z., Wei, L., Chi, E.: Sampling-bias-corrected neural modeling for large corpus item recom- mendations. In: Proceedings of the 13th ACM conference on recommender systems. pp. 269–277 (2019)

    2019

  32. [40]

    arXiv preprint arXiv:2303.18223 (2023)

    Zhao, W.X., Zhou, K., Li, J., Tang, T., Wang, X., Hou, Y., Min, Y., Zhang, B., Zhang, J., Dong, Z., et al.: A survey of large language models. arXiv preprint arXiv:2303.18223 (2023)

    2023 arXiv

  33. [41]

    arXiv preprint arXiv:2302.09419 (2023)

    Zhou, C., Li, Q., Li, C., Yu, J., Liu, Y., Wang, G., Zhang, K., Ji, C., Yan, Q., He, L., et al.: A comprehensive survey on pretrained foundation models: A history from bert to chatgpt. arXiv preprint arXiv:2302.09419 (2023)

    2023 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.