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
打造新電商搜尋體驗- 搜尋意圖辨識
Search
LINE Developers Taiwan
PRO
August 28, 2025
0
20
打造新電商搜尋體驗- 搜尋意圖辨識
LINE Developers Taiwan
PRO
August 28, 2025
Tweet
Share
More Decks by LINE Developers Taiwan
See All by LINE Developers Taiwan
清大企業參訪- Ben
line_developers_tw
PRO
0
800
LLM 商品規格萃取大冒險- Vila
line_developers_tw
PRO
0
930
Playwright/MCP/AI -Winter
line_developers_tw
PRO
0
940
LINE EC Product Catalog Development- Rei
line_developers_tw
PRO
0
930
LINE 與 AI 機器人技術應用現況
line_developers_tw
PRO
0
8
QA Testing
line_developers_tw
PRO
0
260
jcconf_datadev_prod
line_developers_tw
PRO
0
9
jcconf_SPM_prod
line_developers_tw
PRO
0
7
jcconf_LINEPay_prod
line_developers_tw
PRO
0
7
Featured
See All Featured
Rebuilding a faster, lazier Slack
samanthasiow
84
9.3k
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
162
15k
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
253
22k
The Web Performance Landscape in 2024 [PerfNow 2024]
tammyeverts
11
930
Optimising Largest Contentful Paint
csswizardry
37
3.5k
Docker and Python
trallard
46
3.7k
Evolution of real-time – Irina Nazarova, EuRuKo, 2024
irinanazarova
9
1k
GraphQLとの向き合い方2022年版
quramy
49
14k
Understanding Cognitive Biases in Performance Measurement
bluesmoon
31
2.7k
A Tale of Four Properties
chriscoyier
162
23k
Why Our Code Smells
bkeepers
PRO
340
57k
A designer walks into a library…
pauljervisheath
210
24k
Transcript
None
打造新電商搜尋體驗 搜尋意圖辨識 Speaker: Dan Chen
01 02 03 04 如何做搜尋意圖 Prompt vs Finetune LLM 驗證
什麼是搜尋意圖 05 總結 CONTENT
Dan Chen LINE Taiwan EC Dev - Data Scientis All
about Recommendation ….
Dan Chen LINE Taiwan EC Dev - Data Scientis Work
Experience Side Project Intern Work - Consultant (Startup / traditional industries) - DS4G (domestic violence / social housing / elderly issues) - TWiDS / WiDS - Publish TF2.0 books
真實的需求與目的 01 什麼是搜尋意圖?
搜尋意圖是… 特點項目文字 特點項目 特點項目文字 特點項目 關鍵字語句背後 用戶的真實需求與目的 提升用戶體驗 提升平均訂單價值 縮短購買決策週期
特點項目文字 特點項目 顧客 油性肌膚 No Results 生日快樂 星座 *https://www.flaticon.com/
過去機器學習到LLM世代 02 如何做搜尋意圖
Unsupervised Supervised 雖然預測速度快, 但不夠精準 需要大量工人智慧
- Requirement 1. Low latency 2. High precision Dictionary-based
該選擇什麼策略 03 Prompt vs Finetune
Prompt Fine-tune - What is ? - Ref: AI Engineering
Published by O'Reilly Media, Inc. *https://www.flaticon.com/
Prompt Fine-tune - When use? - Ref: AI Engineering Published
by O'Reilly Media, Inc. *https://www.flaticon.com/
如何驗證LLM 04 LLM驗證
*https://www.flaticon.com/ 驗證方法 1. Weak Supervise 2. Cross validation with different
model 3. Self-consistency/Ensemble 4. Synthetic Data
*https://www.flaticon.com/ 開源Performance比較 Breeze Taiwan llama gpt-OSS
*https://www.flaticon.com/ 建立 Reliable Sample
05 Conclusion
None
None