-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpresentation.qmd
214 lines (137 loc) · 6.61 KB
/
presentation.qmd
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
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
---
title: "ChatGPT: AI tools, applications and implications"
author: "Ali Abbas"
format:
revealjs:
slide-number: true
incremental: true
width: 1920
height: 1080
embed-resources: false
preview-links: true
logo: "https://www.mrc-epid.cam.ac.uk/wp-content/uploads/2021/10/colour-epi-logo.png"
footer: "Access at [GitHub Pages](https://usr110.github.io/chatgpt_presentation/presentation.html)"
css: ["theme/theme.css"]
theme: [dark, theme/custom.scss]
html:
code-copy: true
editor: source
execute:
echo: true
warning: false
message: false
size: "huge"
bibliography: references.bib
---
# Agenda
- Intro to ChatGPT
- State of the art
- Examples
- Tools using OpenAI models like ChatGPT (or similar) models
- Issues
- Summary
::: {.notes}
Not an expert so talking as someone with an interest in this area
Brief overview of OpenAI and ChatGPT
Some examples using OpenAI models - like GPT 3
Might take you on a tour of an AI tool
Issues and summary at the end
:::
# ChatGPT
## Overview
- [ChatGPT: Optimizing Language Models for Dialogue](https://chat.openai.com)
- Created by [OpenAI](https://openai.com)
- Debut in November 2022
- Viral due to usability and accessibility
- Built on GPT-3.5 series with Reinforcement Learning from Human Feedback
- Trained on a dataset of over 40GB of text data
- With 175 billion parameters
## (Some of) Applications
- Text generation
- Coherent, creative on any topic and in specific style
- Language translation
- Translation, idiomatic expressions and remembers context
- Dialogue systems
- Conversational agents such as chatbots
- Summarization
- Long to short summary and extract key information
# State of the Art
## Taxonomy AI models
![State of the art - popular generative models @gozalo-brizuela2023](img/state-of-art-gen-AI.png)
## Companies
![AI companies @gozalo-brizuela2023](img/state-of-art-gen-AI-comps.png)
## Timeline
![Timeline @gozalo-brizuela2023](img/state-of-art-gen-AI-timeline.png)
# Examples
## Note to neighbour
![](img/example-chatgpt-site.png)
## Poem
![Poem on benefits of Walking and Cycling](img/chatgpt-poem-cycling-walking.png)
## [Perplexity](https://www.perplexity.ai/)
GPT based search engine
![Children's exposure to PM2.5 affecting their health](img/perplexity-pm2.5.png)
## [Elicit](https://www.elicit.org)
GPT-3 powered AI tool for Literature Review
<video data-autoplay src="vid/elicit.mp4" width="100%" height="85%">
</video>
<!-- {{< (Elicit)[https://www.elicit.org] https://youtu.be/wo9vZccmqwc width="100%" height="85%" >}} -->
# [TutorAI](https://www.tutorai.me/)
OpenAI powered AI tool for creating educational content
## [![TutorAI: Educational content creator](img/tutorAI-twitter.png)](https://twitter.com/DannyHabibs/status/1598069511369867264)
## ![TutorAI example - Physical Activity Epidemiology 1](img/tutorAI-ex1.png)
## ![TutorAI example - Physical Activity Epidemiology 2](img/tutorAI-ex2.png)
## [RTutor](rtutor.ai/)
OpenAI's powerful text-davinci-003 language model to translate natural language into R/Python code
[![](img/rtutor.png){fig-alt="OpenAI's powerful text-davinci-003 language model to translate natural language into R code"}](https://rtutor.ai/)
# [GPT Tools](https://jameshwade.github.io/gpttools/)
## Comment code
<video data-autoplay src="https://user-images.githubusercontent.com/6314313/209890944-3d6a00fa-2d8c-4df7-8a11-f5a5ec3a1391.mov" width="100%" height="85%">
[![Comment Code]()](https://jameshwade.github.io/gpttools/)
## Convert a Script into Functions
<video data-autoplay src="https://user-images.githubusercontent.com/6314313/209890949-4da2bdd7-bcac-4769-9b11-7759b4abb760.mov" width = "100%" height="85%">
<!-- [Convert](https://user-images.githubusercontent.com/6314313/209890949-4da2bdd7-bcac-4769-9b11-7759b4abb760.mov)\](https://jameshwade.github.io/gpttools/) -->
## [GPT Studio](https://github.com/MichelNivard/gptstudio)
Add-on for RStudio for incorportaing OpenAI into project workflows
<video data-autoplay src="vid/gptstudio-chatting-example.mov" width="100%" height="85%">
# Code
## OpenAI R library
<image src="img/open-AI-package.png" width="100%" height="85%">
## Example
Specify a string to create an image
```{r}
#| code-line-numbers: 4
library(openai)
library(magick)
img <- openai::create_image("simplicity")
print(image_read(img$data$url))
```
# Directory of AI tools
# Futurepedia
An AI tools library
<image src="img/futurepedia.png" width="100%" height="85%">
# Issues and Summary
## Issues
- Easier access and universal applicability
- Black box model
- Limited knowledge on how they work
- When they fail
- Capabilities due to emergent properties - anxiety about unanticipated consequences
- Context is not known either
- Downstream bias: intrinsic bias, representational bias, underrepresented and over-represented (Anglocentric population) @bommasani2021
- Compounds existing inequities by producing unfair outcomes, entrenching systems of power, and disproportionately distributing negative consequences of technology to those already marginalized @bommasani2021
- (Almost) exclusively build by industry (Google, Facebook, Microsoft, or Huawei, or startups such as OpenAI) @bommasani2021
- Nefarious purpose: fake news/propagandas, misinformation or any other harmful purpose
## Summary
- Useful in assistance but with almost no transparency or reproducibility, so hard to justify in an academic setting
- Needs to be audited before wider use by academia or general public - legislation perhaps?
- [Rumour has it Google is lobbying DC to shut down other AI efforts, by scaremongering about how unsafe they are](https://twitter.com/balajis/status/1617969414468694017)
- Need a similar but more transparent and less power/computational model
- Among others, Alan Turing Institute is already working on it
- Unless careful attention is applied, it may produce erroneous or even dangerous outputs
## References / further reading
::: nonincremental
- [Futurepedia: AI Tools directory](https://www.futurepedia.io/)
- Chen, M., Tworek, J., Jun, H., Yuan, Q., Pinto, H. P. D. O., Kaplan, J., \... & Zaremba, W. (2021). Evaluating large language models trained on code. *arXiv preprint arXiv:2107.03374*.
- Guo, B., Zhang, X., Wang, Z., Jiang, M., Nie, J., Ding, Y., \... & Wu, Y. (2023). How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detection. *arXiv preprint arXiv:2301.07597*.
- Sobania, D., Briesch, M., Hanna, C., & Petke, J. (2023). An Analysis of the Automatic Bug Fixing Performance of ChatGPT. *arXiv preprint arXiv:2301.08653*.
:::