Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
ACM RecSys 2012: Recommender Systems, Today
Search
Data Science London
October 10, 2012
Technology
2
1.9k
ACM RecSys 2012: Recommender Systems, Today
Neal Lathia @Cambridge_Uni talk at @ds_dln #strataconf 02/10/12
Data Science London
October 10, 2012
Tweet
Share
More Decks by Data Science London
See All by Data Science London
Semi-Supervised Anomaly Detection
datasciencelondon
0
870
Hacking the Rail: Ingesting, analysing & visualising realtime streaming data
datasciencelondon
1
47k
Stateful Data-Parallel Processing
datasciencelondon
0
47k
Semantic web warmed up: Ontologies for the IoT
datasciencelondon
0
110
IoT data ingestion pipelines and Clojure transducers
datasciencelondon
0
250
TrendCalculus: A data science for trends
datasciencelondon
1
48k
Data Science in Mobile Health
datasciencelondon
1
8.3k
Large-scale Recommender Systems on Just a PC (with GraphChi)
datasciencelondon
1
17k
Taming Graph Dynamics at Scale
datasciencelondon
0
8.1k
Other Decks in Technology
See All in Technology
小さな開発会社がWebサービスを作る理由
polidog
PRO
1
180
Google Cloud Next '24でブログを10本書いた方法と勉強会を沸かせた方法
yasumuusan
0
230
JSON攻略法.pdf
miyakemito
8
4.3k
SREとその組織類型
tatsuo48
9
1.6k
コンテナセキュリティの基本と脅威への対策
kyohmizu
3
740
最近たまに見かけるTiDBってなんだ? - Findy
pingcap0315
2
730
ユーザーストーリーのレビューを自動化したみたの
bun913
1
380
VS CodeでAWSを操作しよう
smt7174
7
1.6k
プロトタイピングによる不確実性の低減 / Reducing Uncertainty through Prototyping
ohbarye
5
370
マルチアカウント環境への発見的統制の導入
ch1aki
1
1.3k
テストプロセスで大事にしていること #jasstnano
makky_tyuyan
0
150
「手動オペレーションに定評がある」と言われた私が心がけていること / phpcon_odawara2024
blue_goheimochi
2
360
Featured
See All Featured
Designing for Performance
lara
601
67k
Intergalactic Javascript Robots from Outer Space
tanoku
266
26k
Building a Scalable Design System with Sketch
lauravandoore
455
32k
Web Components: a chance to create the future
zenorocha
305
41k
Typedesign – Prime Four
hannesfritz
36
2.1k
ParisWeb 2013: Learning to Love: Crash Course in Emotional UX Design
dotmariusz
104
6.6k
How to Ace a Technical Interview
jacobian
272
22k
CSS Pre-Processors: Stylus, Less & Sass
bermonpainter
352
28k
Understanding Cognitive Biases in Performance Measurement
bluesmoon
6
990
YesSQL, Process and Tooling at Scale
rocio
163
13k
Building a Modern Day E-commerce SEO Strategy
aleyda
16
6.4k
Practical Orchestrator
shlominoach
181
9.7k
Transcript
acm recsys 2012: recommender systems, today @neal_lathia
warning: daunting task lookout for twitter handles
why #recsys? information overload mailing lists; usenet news (1992) see:
@jkonstan, @presnick
why #recsys? information overload filter failure movies; books; music (~1995)
why #recsys? information overload filter failure creating value advertising; engagement;
connection (today)
@dtunkelang
(1) collaborative “based on the premise that people looking for
information should be able to make use of what others have already found and evaluated” (maltz & ehrlick)
(2) query-less “in September 2010 Schmidt said that one day
the combination of cloud computing and mobile phones would allow Google to pass on information to users without them even typing in search queries”
(3) discovery engines “we are leaving the age of information
and entering the age of recommendation” (anderson)
None
None
input: ratings, clicks, views users → items process: SVD, kNN,
RBM, etc. f(user, item) → prediction ~ rating output: prediction-ranked recommendations measure: |prediction – rating| (prediction – rating)2
traditional problems accuracy, scalability, distributed computation, similarity, cold-start, … (don't
reinvent the wheel)
acm recsys 2012: 5 open problems
problem 1: predictions temporality, multiple co-occurring objectives: diversity, novelty, freshness,
serendipity, explainability
None
problem 2: algorithms more algorithms vs. more data vs. more
rating effort
what is your algorithm doing? f(user, item) → R f(user,
item 1 , item 2 ) → R f(user, [item 1 ...item n ]) → R e.g., @alexk_z @abellogin
problem 3: users + ratings signals, context, groups, intents, interfaces
@xamat
problem 4: items lifestyle, behaviours, decisions, processes, software development
@presnick
problem 5: measurement ranking metrics vs. usability testing vs. A/B
testing
Online Controlled Experiments: Introduction, Learnings, and Humbling Statistics http://www.exp-platform.com/Pages/2012RecSys.aspx
3 key lessons
lesson 1: #recsys is an ensemble ...of disciplines statistics, machine
learning, human-computer interaction, social network analysis, psychology
lesson 2: return to the domain user effort, generative models,
cost of a freakommendation, value you seek to create
@plamere
lesson 3: join the #recsys community learn, build, research, deploy:
@MyMediaLite, @LensKitRS @zenogantner, @elehack contribute, read: #recsyswiki, @alansaid
recommender systems, today @neal_lathia