Master’s Thesis Presentation • Artificial Intelligence — Neural Text Generation from Structured and Unstructured DataExport this event to calendar

Wednesday, August 21, 2019 — 3:00 PM EDT

Hamidreza Shahidi, Master’s candidate
David R. Cheriton School of Computer Science

A number of researchers have recently questioned the necessity of increasingly complex neural network (NN) architectures. In particular, several recent papers have shown that simpler, properly tuned models are at least competitive across several natural language processing (NLP) tasks. 

In this thesis, we show that this is also the case for text generation from structured and unstructured data. Specifically, we consider table-to-text generation and neural question generation (NQG) tasks for text generation from structured and unstructured data respectively. Table-to-text generation aims to generate a description based on a given table, and NQG is the task of generating a question from a given passage where the generated question can be answered by a certain sub-span of the passage using NN models. Experiments demonstrate that a basic attention-based seq2seq model trained with exponential moving average technique achieves state of the art in both tasks. 

We further investigate using reinforcement learning (RL) with different reward functions to refine our pre-trained model for both tasks.

Location 
DC - William G. Davis Computer Research Centre
2310
200 University Avenue West

Waterloo, ON N2L 3G1
Canada

S M T W T F S
28
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
29
30
31
  1. 2019 (180)
    1. September (7)
    2. August (18)
    3. July (12)
    4. June (23)
    5. May (23)
    6. April (32)
    7. March (25)
    8. February (16)
    9. January (24)
  2. 2018 (221)
    1. December (16)
    2. November (19)
    3. October (26)
    4. September (23)
    5. August (17)
    6. July (20)
    7. June (13)
    8. May (25)
    9. April (34)
    10. March (24)
    11. February (3)
    12. January (1)
  3. 2017 (36)
  4. 2016 (21)
  5. 2015 (36)
  6. 2014 (33)
  7. 2013 (23)
  8. 2012 (4)
  9. 2011 (1)
  10. 2010 (1)
  11. 2009 (1)
  12. 2008 (1)