Gravity R&D (full name: Gravity Research & Development Zrt.) is an IT vendor specialized in recommender systems. Gravity was founded by members of the Netflix Prize team "Gravity".
| Company type | Private |
|---|---|
| Industry | Software |
| Founded | 2009 |
| Headquarters | , |
Area served | Worldwide |
Key people | Domonkos Tikk (CEO & co-founder) Bottyán Németh (product owner, co-founder) |
| Products | Yusp, Yuspify for e-commerce |
| Services | IT services, personalization, SaaS |
| Owner | Hungarian institutional strategic investors, Wojciech Uzdelewicz,[1] founders |
Number of employees | 25 |
| Website | yusp.com |
Gravity is headquartered in Hungary (Budapest & Győr) with a subsidiary in Japan.
History
editNetflix Prize
editThe Netflix Prize was an open competition for the best collaborative filtering algorithm to predict user ratings for films, based on previous ratings. The prize would be awarded to the team achieving over 10% improvement over Netflix's own Cinematch algorithm.
The team "Gravity" was the front runner during January—May 2007.[2]
The leading position was achieved again in October 2007 in collaboration with the team "Dinosaur Planet" under the name "When Gravity and Dinosaurs Unite".
In January 2009, the two teams founded "Grand Prize Team" to initiate even wider collaboration that resulted in being one of the leading teams throughout 2009.
On July 25th 2009, the team "The Ensemble", a merger of the teams "Grand Prize Team" and "Opera Solutions and Vandelay United", achieved a 10.10% improvement over Cinematch on the Quiz set.[3]
On September 18, 2009, Netflix announced team "BellKor's Pragmatic Chaos" as the prize winner, and the prize was awarded to the team in a ceremony on September 21, 2009.[4] "The Ensemble" team had in fact succeeded to match the winning "BellKor" team's result, but since "BellKor" submitted their results 20 minutes earlier, the rules award the prize to them.[5][6]
Details on the algorithms developed by the Gravity team can be found in their scientific publications.[7][8][9] Some algorithms are patented in the US.[10]
The data mining team of the company is actively doing research in the field of recommender systems and publish their recent results regularly.[11][12][13][14][15][16][17][18]
Yusp
editOn the model of P&G, Gravity separated company name and product name in 2017. The company name will remain Gravity while the brand name is changed to Yusp. Yusp is the name of the new generation personalization engine. Under Yusp, Gravity currently develops different product lines for enterprise, online-only, and bricks and mortar retail, telecommunications and retail banking customers and potential customers.
References
edit- ^ "How Hedge Funds Rate Wall Street Analysts, 2003".
- ^ Hafner, Katie (June 4, 2007). "Netflix Prize Still Awaits a Movie Seer". The New York Times. Retrieved 2010-03-07.
- ^ "The Ensemble". 2009-07-25.
- ^ "Grand Prize awarded to team BellKor's Pragmatic Chaos". Netflix Prize Forum. 2009-09-21. Archived from the original on 2012-05-07. Retrieved 2012-05-07.
- ^ Steve Lohr (2009-09-21). "A $1 Million Research Bargain for Netflix, and Maybe a Model for Others". New York Times.
- ^ "Mátrixfaktorizáció egymillió dollárért". Index. 2009-08-07.
- ^ Takács, G. B.; Pilászy, I. N.; Németh, B. N.; Tikk, D. (2007). "Major components of the gravity recommendation system". ACM SIGKDD Explorations Newsletter. 9 (2): 80. doi:10.1145/1345448.1345466. S2CID 4518283.
- ^ Gábor Takács; István Pilászy; Bottyán Németh; Domonkos Tikk (2007), "On the Gravity Recommendation System" (PDF), in Gábor Takács; István Pilászy; Bottyán Németh and Domonkos Tikk (eds.), Proc. KDD Cup Workshop at SIGKDD, San Jose, California, pp. 22–30, retrieved 2010-04-15
{{citation}}: CS1 maint: location missing publisher (link) - ^ Gábor Takács; István Pilászy; Bottyán Németh; Domonkos Tikk (2009), Scalable Collaborative Filtering Approaches for Large Recommender Systems (PDF)
- ^ US patent 8676736, Pilaszy, et al., "Recommender systems and methods using modified alternating least squares algorithm", issued 2014-03-18
- ^ István Pilászy; Domonkos Tikk (2009), "Recommending new movies: Even a few ratings are more valuable than metadata", Proceedings of the third ACM conference on Recommender systems, RecSys '09, pp. 93–100, doi:10.1145/1639714.1639731, ISBN 978-1-60558-435-5, S2CID 17687390
- ^ István Pilászy; Dávid Zibriczky; Domonkos Tikk (2010), "Fast ALS-based matrix factorization for explicit and implicit feedback datasets", Proceedings of the fourth ACM conference on Recommender systems - Rec Sys '10, RecSys '10, pp. 71–78, doi:10.1145/1864708.1864726, ISBN 978-1-60558-906-0, S2CID 1816937
- ^ Gábor Takács; István Pilászy; Domonkos Tikk (2011), "Applications of the conjugate gradient method for implicit feedback collaborative filtering", Proceedings of the fifth ACM conference on Recommender systems - Rec Sys '11, RecSys '11, pp. 297–300, doi:10.1145/2043932.2043987, ISBN 978-1-4503-0683-6, S2CID 3342766
- ^ Balázs Hidasi; Domonkos Tikk (2012), "Fast ALS-Based Tensor Factorization for Context-Aware Recommendation from Implicit Feedback", Machine Learning and Knowledge Discovery in Databases, Lecture Notes in Computer Science, vol. 7524, pp. 67–82, arXiv:1204.1259, doi:10.1007/978-3-642-33486-3_5, ISBN 978-3-642-33485-6, S2CID 9480129
- ^ Gábor Takács; Domonkos Tikk (2012), "Alternating least squares for personalized ranking", Proceedings of the sixth ACM conference on Recommender systems - Rec Sys '12, RecSys '12, pp. 83–90, doi:10.1145/2365952.2365972, ISBN 978-1-4503-1270-7, S2CID 3357762
- ^ Balázs Hidasi; Domonkos Tikk (2013), "Context-aware item-to-item recommendation within the factorization framework", Proceedings of the 3rd Workshop on Context-awareness in Retrieval and Recommendation - CaRR '13, pp. 19–25, doi:10.1145/2442670.2442675, ISBN 978-1-4503-1847-1, S2CID 14906053
- ^ Alan Said; Domonkos Tikk; Paolo Cremonesi (2014), "Benchmarking: A Methodology for Ensuring the Relative Quality of Recommendation Systems in Software Engineering", Recommendation Systems in Software Engineering, pp. 275–300, doi:10.1007/978-3-642-45135-5_11, hdl:11311/1006649, ISBN 978-3-642-45134-8, S2CID 38607259
- ^ Balázs Hidasi; Domonkos Tikk (2014), "Approximate modeling of continuous context in factorization algorithms", Proceedings of the 4th Workshop on Context-Awareness in Retrieval and Recommendation, pp. 3–9, doi:10.1145/2601301.2601303, ISBN 9781450327237, S2CID 17842678