Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
06 Models for Use-Cases
Search
Sponsored
·
Your Podcast. Everywhere. Effortlessly.
Share. Educate. Inspire. Entertain. You do you. We'll handle the rest.
→
LiLa'16
March 20, 2016
Research
65
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
06 Models for Use-Cases
LiLa'16
March 20, 2016
More Decks by LiLa'16
See All by LiLa'16
01 Introduction
lila16ecir
0
89
02 Online Evaluation
lila16ecir
0
95
03 LL4IR Architecture
lila16ecir
0
75
04 Use-Cases
lila16ecir
0
79
05 API Interactions I.
lila16ecir
0
69
07 API Interactions II.
lila16ecir
0
61
08 Interpreting Feedback
lila16ecir
0
110
09 API Interactions III.
lila16ecir
0
62
10 Simulations
lila16ecir
0
87
Other Decks in Research
See All in Research
NLP colloquium: AI Safety Survey
kanekomasahiro
2
1.1k
PGDM: Physically Guided Diffusion Model for L Downscaling
satai
3
470
ふとした出会いで生まれたSkillが、 社内利用1位になるまで
mikimhk
21
24k
Source Code Diff Revolution
tsantalis
0
140
PHTalks Bengaluru - SSRF When All Else Fails
dk999
0
1.1k
敵対生成プロンプト同時探索による内省型プロンプト最適化
kinoue_smarthr
0
410
医療LLMの現在地〜最新研究から社会実装までを考える〜
kento1109
1
1.7k
シングルチャネルマルチトーカー音声認識の進展
ryomasumura
0
260
LA-Bench 2025:実験指示から実行可能手順を生成するためのデータセット/LA-Bench 2025: A Dataset for Generating Executable Experimental Procedures from Experimental Instructions
stktu
0
150
CDCL を用いた MILP の厳密解法
imai448
0
250
「AIとWhyを深堀る」をAIと深堀る
iflection
0
610
横浜市長(山中氏)の言動にかかる第三者による調査報告書
y150saya
0
230
Featured
See All Featured
Ten Tips & Tricks for a 🌱 transition
stuffmc
0
210
Evolving SEO for Evolving Search Engines
ryanjones
0
280
Google's AI Overviews - The New Search
badams
0
1.6k
Reflections from 52 weeks, 52 projects
jeffersonlam
356
21k
DevOps and Value Stream Thinking: Enabling flow, efficiency and business value
helenjbeal
1
380
How Software Deployment tools have changed in the past 20 years
geshan
1
34k
The Illustrated Children's Guide to Kubernetes
chrisshort
51
53k
Raft: Consensus for Rubyists
vanstee
141
7.7k
Introduction to Domain-Driven Design and Collaborative software design
baasie
1
980
BBQ
matthewcrist
89
10k
GraphQLとの向き合い方2022年版
quramy
50
15k
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
162
16k
Transcript
Anne Schuth (Blendle / University of Amsterdam, The Netherlands) Krisztian
Balog (University of Stavanger, Norway) Tutorial at ECIR 2016 in Padua, Italy Models for Use-Cases
Overview • Two use-cases • Product search • Academic literature
search • Simple shared approach • Fielded document representation
Baseline model • Single-field document retrieval • E.g., BM25 score
( d, q ) = X t2q ft,d · (1 + k1) ft,d + k1(1 b + b |d| avgdl ) · idft
Single-field document represenation • Using only product name field •
Using catchall field docid R-d3504 product_name LEGO DUPLO Kreatív láda 10817 docid R-d3504 catchall LEGO DUPLO Kreatív láda 10817 Építőjáték, LEGO LEGO Éld ki kreativitásod minden évszakban LEGO DUPLO elemekkel!
Fielded extension • Multi-field document retrieval, e.g., BM25F score (
d, q ) = X t2q ˜ ft,d k1 + ˜ ft,d · idft ˜ ft,d = X i wi ft,di Bi Combining term frequencies across fields Field weight Soft normalization for field i
Multi-field document representation docid R-d3504 product_name LEGO DUPLO Kreatív láda
10817 brand LEGO DUPLO short_description Éld ki kreativitásod minden évszakban LEGO DUPLO elemekkel! catchall LEGO DUPLO Kreatív láda 10817 Építőjáték, LEGO LEGO Éld ki kreativitásod minden évszakban LEGO DUPLO elemekkel!
Incorporating historical feedback • Combining historical score with retrieval score
• A simple approach: • or