Upgrade to Pro — share decks privately, control downloads, hide ads and more …

An_Embarrassingly_Simple_Approach_for_Transfer_...

MARUYAMA
July 10, 2019
0
130

 An_Embarrassingly_Simple_Approach_for_Transfer_Learning_from_Pretrained_Language_Models_.pdf

MARUYAMA

July 10, 2019
Tweet

More Decks by MARUYAMA

See All by MARUYAMA
vampire.pdf
tmaru0204
0
140
Misspelling_Oblivious_Word_Embedding.pdf
tmaru0204
0
170
Simple_Unsupervised_Summarization_by_Contextual_Matching.pdf
tmaru0204
0
150
Controlling_Text_Complexity_in_Neural_Machine_Translation.pdf
tmaru0204
0
140
20191028_literature-review.pdf
tmaru0204
0
130
Hint-Based_Training_for_Non-Autoregressive_Machine_Translation.pdf
tmaru0204
0
120
Soft_Contextual_Data_Augmentation_for_Neural_Machine_Translation_.pdf
tmaru0204
0
140
Addressing_Trobulesome_Words_in_Neural_Machine_Translation.pdf
tmaru0204
0
130
Simple_Unsupervised_Keyphrase_Extraction_using_Sentence_Embeddings.pdf
tmaru0204
0
170

Featured

See All Featured
Design and Strategy: How to Deal with People Who Don’t "Get" Design
morganepeng
127
18k
CoffeeScript is Beautiful & I Never Want to Write Plain JavaScript Again
sstephenson
160
15k
Visualization
eitanlees
146
15k
How STYLIGHT went responsive
nonsquared
96
5.3k
Automating Front-end Workflow
addyosmani
1366
200k
Java REST API Framework Comparison - PWX 2021
mraible
28
8.3k
For a Future-Friendly Web
brad_frost
176
9.5k
[RailsConf 2023] Rails as a piece of cake
palkan
53
5.1k
Embracing the Ebb and Flow
colly
84
4.5k
Designing for Performance
lara
604
68k
Fantastic passwords and where to find them - at NoRuKo
philnash
50
2.9k
What's in a price? How to price your products and services
michaelherold
244
12k

Transcript

  1. "O&NCBSSBTTJOHMZ4JNQMF"QQSPBDIGPS 5SBOTGFS-FBSOJOHGSPN1SFUSBJOFE-BOHVBHF.PEFMT "MFYBOESB$ISPOPQPVMPV $ISJTUPT#B[JPUJT "MFYBOESPT1PUBNJBOPT /""$-)-5 QBHFTr -JUFSBUVSFSFWJFX /BHBPLB6OJWFSTJUZPG5FDIOPMPHZ5BLVNJ.BSVZBNB

  2. "CTUSBDU ⾣FOEUPFOEʹֶशͤ͞ΔγϯϓϧͳసҠֶशख๏ΛఏҊ ⾣ࣄલֶशͨ͠ݴޠϞσϧΛ෼ྨλεΫʹpOFUVOJOH ⾣༷ʑͳςΩετ෼ྨλεΫͰɺطଘख๏Λ௒͑ΔੑೳΛୡ੒

  3. 3FMBUFEXPSL ⾣6-.'J5 %JTDSJNJOBUJWFpOFUVOJOH֤૚Ͱֶश཰ΛมԽͤ͞Δ 4MBOUFEUSJBOHVMBSMFBSOJOHSBUFT

  4. .PEFM 4J"5- ⾣-.1SFUSBJOJOH ⾣5SBOTGFSBVYJMJBSZMPTT ɾFYQPOFOUJBMEFDBZPGЍ FQPDIຖʹЍΛࢦ਺తʹݮਰ ɾTFRVFOUJBMVOGSFF[JOH ্ͷ૚͔Βॱ࣍ղౚ͠ɺpOFUVOJOH

  5. &YQFSJNFOUBMTFUVQT ⾣%BUBTFUT ɾ*SPOZJSPOZEFUFDUJPO ɾ4FOUTFOUJNFOUBOBMZTJT ɾ4$W 4$WTBSDBTNEFUFDUJPO ɾ1TZDI&YQFNPUJPOSFDPHOJUJPO

  6. 3FTVMUT ⾣"CMBUJPOTUVEZ

  7. 3FTVMUT ⾣/VNCFSPGFYBNQMFT ɾ4J"5- DPOUJOVPVTMJOFT ɾ6-.'J5 EBTIFEMJOFT dσʔλఔ౓Ͱɺ 6-.'J5ʹඖఢ͢Δੑೳ

  8. 3FTVMUT ⾣&YQPOFOUJBMEFDBZPGЍ *OJUJBMWBMVF 'JOBMWBMVF

  9. $PODMVTJPO ⾣FOEUPFOEʹֶशͤ͞ΔγϯϓϧͳసҠֶशख๏ΛఏҊ ⾣ࣄલֶशͨ͠ݴޠϞσϧΛ෼ྨλεΫʹpOFUVOJOH ⾣༷ʑͳςΩετ෼ྨλεΫͰɺطଘख๏Λ௒͑ΔੑೳΛୡ੒