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
サブセット探索を用いた高速なkNNニューラル機械翻訳
Search
Hiroyuki Deguchi
March 22, 2024
Research
0
140
サブセット探索を用いた高速なkNNニューラル機械翻訳
第8回AAMTセミナー
AAMT若手翻訳研究会
最優秀賞
Hiroyuki Deguchi
March 22, 2024
Tweet
Share
More Decks by Hiroyuki Deguchi
See All by Hiroyuki Deguchi
20250226 NLP colloquium: "SoftMatcha: 10億単語規模コーパス検索のための柔らかくも高速なパターンマッチャー"
de9uch1
0
570
20240820: Minimum Bayes Risk Decoding for High-Quality Text Generation Beyond High-Probability Text
de9uch1
0
280
20240226_AAMT-Japio
de9uch1
0
160
Searching for Needles in a Haystack: On the Role of Incidental Bilingualism in PaLM’s Translation Capability
de9uch1
0
130
Paper Reading: Sampling-Based Approximations to Minimum Bayes Risk Decoding for Neural Machine Translation
de9uch1
0
180
My Research Environmental Setup
de9uch1
0
300
Nearest Neighbor Machine Translation
de9uch1
0
250
Paper Reading - Dynamic Programming Encoding for Subword Segmentation in Neural Machine Translation
de9uch1
0
280
paper reading - Tree Transformer
de9uch1
0
250
Other Decks in Research
See All in Research
A scalable, annual aboveground biomass product for monitoring carbon impacts of ecosystem restoration projects
satai
4
350
Nullspace MPC
mizuhoaoki
1
200
Creation and environmental applications of 15-year daily inundation and vegetation maps for Siberia by integrating satellite and meteorological datasets
satai
3
400
MetaEarth: A Generative Foundation Model for Global-Scale Remote Sensing Image Generation
satai
4
320
Time to Cash: The Full Stack Breakdown of Modern ATM Attacks
ratatata
0
160
Towards a More Efficient Reasoning LLM: AIMO2 Solution Summary and Introduction to Fast-Math Models
analokmaus
2
910
機械学習と数理最適化の融合 (MOAI) による革新
mickey_kubo
1
390
AIスパコン「さくらONE」のLLM学習ベンチマークによる性能評価 / SAKURAONE LLM Training Benchmarking
yuukit
2
690
電通総研の生成AI・エージェントの取り組みエンジニアリング業務向けAI活用事例紹介
isidaitc
1
1k
生成的推薦の人気バイアスの分析:暗記の観点から / JSAI2025
upura
0
280
cvpaper.challenge 10年の軌跡 / cvpaper.challenge a decade-long journey
gatheluck
3
350
Submeter-level land cover mapping of Japan
satai
3
420
Featured
See All Featured
Designing Dashboards & Data Visualisations in Web Apps
destraynor
231
53k
How to Create Impact in a Changing Tech Landscape [PerfNow 2023]
tammyeverts
55
3k
Fireside Chat
paigeccino
40
3.7k
Building Better People: How to give real-time feedback that sticks.
wjessup
369
20k
How to train your dragon (web standard)
notwaldorf
97
6.3k
Bootstrapping a Software Product
garrettdimon
PRO
307
110k
Build The Right Thing And Hit Your Dates
maggiecrowley
37
2.9k
BBQ
matthewcrist
89
9.8k
Build your cross-platform service in a week with App Engine
jlugia
232
18k
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
35
6.1k
Rails Girls Zürich Keynote
gr2m
95
14k
A better future with KSS
kneath
239
18k
Transcript
𝒌
◼ ⚫ ⚫ ◼ ⚫ (Zhang+, NAACL2018; Gu+, AAAI2018; Khandelwal+,
ICLR2021) ▶ (Nagao, 1984) ▶ ⚫ 𝑘 (Khandelwal+, ICLR2021) ▶ ▶ ▶ Guiding Neural Machine Translation with Retrieved Translation Pieces (Zhang+, NAACL2018) Search Engine Guided Neural Machine Translation (Gu+, AAAI2018) Nearest Neighbor Machine Translation (Khandelwal+, ICLR2021) A framework for a mechanical translation between Japanese and English by analogy principle (Nagao, 1984)
◼ ◼ ⚫ ⚫
𝒌 (Khandelwal+, ICLR2021) ◼ ⚫ ⚫ ⚫ ◼ ⚫ ▶
⚫ ▶ ≈ Nearest Neighbor Machine Translation (Khandelwal+, ICLR2021) 𝒙 𝒚
𝒌 (Khandelwal+, ICLR2021) 𝒌𝑖 ∈ ℝ𝐷 𝑓 𝒙, 𝒚<𝑡 ∈
ℝ𝐷 Nearest Neighbor Machine Translation (Khandelwal+, ICLR2021) ◼ 𝑘 ◼ ⚫ ⚫ 𝑝𝑘NN 𝑦𝑡 𝒙, 𝒚<𝑡 ∝ 𝑖=1 𝑘 𝟙𝑦𝑡=𝑣𝑖 exp − 𝒌𝑖 − 𝑓 𝒙, 𝒚<𝑡 2 2 𝜏 ◼ 𝑘
𝒌 ◼ (Martins+, EMNLP2022) ◼ (Meng+, ACLFindings2022) ⚫ 𝑘 𝑘
𝜆 = 0.5 𝑘 = 16 Chunk-based Nearest Neighbor Machine Translation (Martins+, EMNLP2022) Fast Nearest Neighbor Machine Translation (Meng+, ACL Findings2022)
𝒌 ◼ 𝑘 ◼ ⚫ 𝑘 (Matsui+, ACMMM2018) ⚫ 𝑘
𝑘 𝑘 Reconfigurable Inverted Index (Matsui+, ACMMM2018) 𝒌
◼ ⚫ 𝑘 ⚫ 𝑘 ◼ ◼ 𝑘
𝑛 𝑘 1 1 1 1 1 1 1 1
1
𝑛 𝑘 1 1 1 1 1 1 1 1
1
𝑛 𝑘 1 1 1 1 1 1 1 1
1
⚫ ⚫ ⚫ ⚫ ⚫ 𝑘 𝜆 = 0.5 𝑘
= 16 𝑛 = 56
𝑘 𝑘 ◼ 𝑘 ⚫ ▶ ⚫ ▶
◼ 𝑘 𝒌 𝒌
◼ ⚫ 𝑘
𝒌 𝒌 ◼ ⚫ ⚫ ◼ 𝑘 ⚫ ⚫ ◼
⚫
⚫ ⚫ ▶ ⚫ ▶