PhD Seminar • Natural Language Processing • Segatron: Segment-aware Transformer for Language ModelingExport this event to calendar

Monday, December 12, 2022 — 10:00 AM to 11:00 AM EST

Please note: This PhD seminar will take place online.

He (Richard) Bai, PhD candidate
David R. Cheriton School of Computer Science

Supervisor: Professor Ming Li

TL;DR: This seminar would cover our AAAI work about LM and BERT pretraining, and our recent unpublished work about GPT-3 13B model pretraining. We show our Segatron consistently outperform BERT and GPT-3 under the same training configuration over various model size and evaluation tasks. We will also talk about our findings on the validation of LLM.

Abstract: Transformers are powerful for sequence modeling. Nearly all state-of-the-art language models and pre-trained language models are based on the Transformer architecture. However, it distinguishes sequential tokens only with the token position index. We hypothesize that better contextual representations can be generated from the Transformer with richer positional information. To verify this, we propose a segment-aware Transformer (Segatron), by replacing the original token position encoding with a combined position encoding of paragraph, sentence, and token.

We first introduce the segment-aware mechanism to Transformer-XL, which is a popular Transformer-based language model with memory extension and relative position encoding. We find that our method can further improve the Transformer-XL base model and large model, achieving 17.1 perplexity on the WikiText-103 dataset. We further investigate the pre-training masked language modeling task with Segatron. Experimental results show that BERT pre-trained with Segatron (SegaBERT) can outperform BERT with vanilla Transformer on various NLP tasks, and outperforms RoBERTa on zero-shot sentence representation learning.

Paper link: https://ojs.aaai.org/index.php/AAAI/article/view/17485


To join this PhD seminar on Zoom, please go to https://uwaterloo.zoom.us/j/9968684700.

Location 
Online PhD seminar
200 University Avenue West

Waterloo, ON N2L 3G1
Canada
Event tags 

S M T W T F S
29
30
31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
1
2
3
4
  1. 2023 (32)
    1. March (1)
    2. February (9)
    3. January (22)
  2. 2022 (245)
    1. December (20)
    2. November (28)
    3. October (15)
    4. September (12)
    5. August (29)
    6. July (23)
    7. June (17)
    8. May (20)
    9. April (24)
    10. March (22)
    11. February (16)
    12. January (19)
  3. 2021 (210)
  4. 2020 (217)
  5. 2019 (255)
  6. 2018 (217)
  7. 2017 (36)
  8. 2016 (21)
  9. 2015 (36)
  10. 2014 (33)
  11. 2013 (23)
  12. 2012 (4)
  13. 2011 (1)
  14. 2010 (1)
  15. 2009 (1)
  16. 2008 (1)