Index: 00:00 Introduction to EASE Events - Duncan Nicholas, EASE President 02:14 Housekeeping 03:45 EASE RUclips & Resources 04:31 Session Introduction Avi Staiman 07:39 GPT and Large Language Models: A Primer - Phill Jones 08:41 Who’s Behind GPT? 10:28 HOW IT WORKS 10:57 Generating Text: The Stochastic Parrot 12:38 Neural Networks allow for a best guess 14:50 Turning text into numbers - Natural Language Processing 15:36 Transformer models are not robots in disguise 16:50 Pre-training to create plausible text 18:38 CAPABILITIES AND LIMITATIONS 19:14 What GPT can do and can’t do 21:00 Garbage in…offensive garbage out 24:16 LLMs can behave in ways that are…unsettling 26:00 CURRENT AND FUTURE DEVELOPMENTS 26:08 Existing and emerging use cases 28:10 It’s still early days 30:09 An evolving technology race 31:32 AI in Research Publishing: Exploring the Great Unknown - Avi Staiman 36:00 Wordle for knowledge 39:22 WHAT ARE RESEARCHERS ALREADY USING CHATGPT FOR? 39:58 A level playing field for ESL scholars 41:23 A cooperative research advisor 43:57 A research assistant 45:31 A personal peer reviewer 46:14 A personal publicist 47:38 ARE ACADEMICS USING IT? 51:13 Q&A
Extra answers to questions raised in the Q&A that we didn’t have time to get to can be found here: docs.google.com/document/d/1h51b3uzrTjmAD_kG3Ka5t935wuF4n8rsaYahEdt0gcw/edit?usp=sharing
Great! But few more questions are left unanswered, can you look at them and get some answers, here is one: "Is there any projects out there using the academic corpus of texts/literature and using it in a healthy way? who could be interested in discussing this forward?"
Index:
00:00 Introduction to EASE Events - Duncan Nicholas, EASE President
02:14 Housekeeping
03:45 EASE RUclips & Resources
04:31 Session Introduction Avi Staiman
07:39 GPT and Large Language Models: A Primer - Phill Jones
08:41 Who’s Behind GPT?
10:28 HOW IT WORKS
10:57 Generating Text: The Stochastic Parrot
12:38 Neural Networks allow for a best guess
14:50 Turning text into numbers - Natural Language Processing
15:36 Transformer models are not robots in disguise
16:50 Pre-training to create plausible text
18:38 CAPABILITIES AND LIMITATIONS
19:14 What GPT can do and can’t do
21:00 Garbage in…offensive garbage out
24:16 LLMs can behave in ways that are…unsettling
26:00 CURRENT AND FUTURE DEVELOPMENTS
26:08 Existing and emerging use cases
28:10 It’s still early days
30:09 An evolving technology race
31:32 AI in Research Publishing: Exploring the Great Unknown - Avi Staiman
36:00 Wordle for knowledge
39:22 WHAT ARE RESEARCHERS ALREADY USING CHATGPT FOR?
39:58 A level playing field for ESL scholars
41:23 A cooperative research advisor
43:57 A research assistant
45:31 A personal peer reviewer
46:14 A personal publicist
47:38 ARE ACADEMICS USING IT?
51:13 Q&A
Extra answers to questions raised in the Q&A that we didn’t have time to get to can be found here: docs.google.com/document/d/1h51b3uzrTjmAD_kG3Ka5t935wuF4n8rsaYahEdt0gcw/edit?usp=sharing
Great! But few more questions are left unanswered, can you look at them and get some answers, here is one: "Is there any projects out there using the academic corpus of texts/literature and using it in a healthy way? who could be interested in discussing this forward?"