Skip to main content

Meeting minutes for 2023-07-13

Please find below the meetings notes for our 2023/07/13 session.

00:39Peter Smulovics (he/him, MS)Welcomes everyone. Good evening. One or 2 more minutes for people to join.
03:17Julia RitterJulia Ritter kicks off the meeting with meeting notices about project guidelines, codes of conduct, industry participation, antitrust policies, and the recording of meetings.
04:50Peter Smulovics (he/him, MS)Peter Smulovics requests attendees to visit the Github link for attendance.
05:03Keith J. O'DonnellKeith introduces the agenda and skips the Poc program part.
05:34Keith J. O'DonnellKeith provides updates on blogs, videos, upcoming events, and introduces new team members Carly Richmond, Leonardo Mordasini, and Polina Levyant.
09:02Keith J. O'DonnellKeith discusses AI, weak AI, strong AI, artificial general intelligence (AGI), and superintelligence. He then delves into the history of AI, including the Turing test and adversarial programs.
14:50Keith J. O'DonnellKeith mentions the XCD and adversarial programs playing checkers in the 1950s.
15:50Keith J. O'Donnellmentions the clever development of the first instance of Alpha beta through a serial network, laying the groundwork for future AI applications.
16:26Keith J. O'Donnelldiscusses the emergence of the term "artificial intelligence" and how people attempted to cheat homework using intelligent programs.
17:00Keith J. O'Donnelltalks about the development of algorithms in applied mathematics and provides a quick break.
16:16Keith J. O'Donnellintroduces an algorithm and challenges the audience to figure out its meaning.
16:16Keith J. O'Donnellconfirms that the algorithm proves one plus one equals two and explains its significance in training algorithms for problem-solving.
17:00Keith J. O'Donnelldescribes the development of perception and the first single-layer neural network by Frank Rosenblatt.
17:00Keith J. O'Donnellmentions the development of chatbots, industrial robots, and advancements in natural language processing in the 1960s.
18:00Keith J. O'Donnellhighlights the significance of Dendr and its impact on organic chemistry and drug development.
18:00Keith J. O'Donnelldiscusses the first general-purpose robot, Shaky, and its ability to reason and make decisions.
18:00Keith J. O'Donnelltalks about the development of backpropagation and its impact on deep learning.
18:00Keith J. O'Donnellmentions the progress made in the 1970s and Marvin Minsky's role in speech recognition.
19:00Keith J. O'Donnellintroduces the Stanford car, the first autonomous vehicle, and its navigation capabilities.
19:00Keith J. O'Donnelltalks about advancements in robotics, including music-reading robots and self-driving cars in the 1980s.
19:00Keith J. O'Donnellmentions the portrayal of technology in movies and its impact on public perception.
20:00Keith J. O'Donnelldiscusses the development of long short-term memory (LSTM) and its applications in handwriting and speech recognition.
20:00Keith J. O'Donnellrecounts the game between Gary Kasparov and Deep Blue, showcasing computers' ability to play complex games.
20:00Keith J. O'Donnelldescribes the Furby as a simulation of artificial intelligence and its impact on familiarizing people with machine learning concepts.
20:00Keith J. O'Donnellmoves into the 21st century, highlighting the advancements in robots working in hospitality and self-driving cars.
21:00Keith J. O'Donnellmentions the triumph of AlphaGo and its ability to defeat human players in the game of Go.
21:00Keith J. O'Donnellprovides an overview of the current landscape of AI development tools and discusses their applications in finance and insurance.
21:00Keith J. O'Donnellmentions the plans to create primers focusing on commercialization, ethics, and security vulnerabilities.
22:00Keith J. O'Donnelltalks about the importance of collaboration and use within the open-source community.
22:00Keith J. O'Donnelldiscusses technology readiness levels and their relevance in assessing the maturity of AI technologies.
23:00Keith J. O'Donnellintroduces the concept of technology readiness scale and its different stages.
24:00Keith J. O'Donnellexplains the purpose of technology readiness levels in evaluating the maturity of AI technologies.
25:00Keith J. O'Donnelldiscusses specific technology readiness levels for different AI disciplines and identifies potential blockers.
26:00Keith J. O'Donnellhighlights the importance of model validation, monitoring, and machine learning deployment platforms.
26:00Keith J. O'Donnelltalks about the importance of resource optimization and the need for specific hardware and operating systems.
27:00Keith J. O'Donnelldiscusses computer vision and natural language processing, emphasizing the need for common frameworks and standards.
28:00Keith J. O'Donnelladdresses the intersection of quantum computing and AI and mentions the Venn diagram overlap of emerging technologies.
28:00Keith J. O'Donnellresponds to a comment about missing aspects in the primer related to use cases, training, and data sources.
29:00Keith J. O'Donnellencourages discussion and comments from the participants.
29:00Keith J. O'Donnellopens the floor for any further comments or questions.
29:26Velvárt Andrásmentions the entire process of developing an AI and raises questions about key participants in the industry and the need for compute power.
29:32Velvárt Andrásasks about copyright and licensing considerations for the primer and its readiness.
30:02Keith J. O'Donnellproposes creating a series of videos to address the mentioned topics and adds them to the list of things to include in the primer.
30:30Nick Williamssuggests including a primer section on definitions of terms and discusses the importance of providing content that doesn't require watching videos.
31:03Keith J. O'Donnellacknowledges the vastness of the field and agrees to include governance and regulatory aspects in the primer.
31:29Keith J. O'Donnellthanks Michael Wilson for his contribution and discusses the value propositions that will be included in the primer.
34:48Nick Williamssuggests adding a primer section on governance and regulations in the financial sector for using AI.
35:54Keith J. O'Donnellmentions including security vulnerabilities, fair use, and licenses in the primer and highlights the need for governance considerations.
36:05Keith J. O'Donnellinvites participants to raise any questions or contribute to the discussion.
36:11Keith J. O'Donnellpresents the core blockers identified for experimentation, including AI chipsets, neural networks, and generative adversarial networks.
38:18Keith J. O'Donnelldiscusses the need to understand the needs of industries and establish a common data language and synthetic data rule set.
39:07Keith J. O'Donnellmentions the importance of working with regulators, audit and risk professionals, and addressing data quality and observability.
39:26Keith J. O'Donnellhighlights the need for experiment tracking and addressing technical debt in AI-specific tasks and platforms.
39:37Keith J. O'Donnellemphasizes the mathematical modeling of drifting bias and the importance of avoiding PR disasters like the "Tay to Twitter" incident.
40:04Keith J. O'Donnellintroduces the goal of creating a machine learning application in a box for micro deployment and experimentation on various platforms.
40:13Keith J. O'Donnelldiscusses the customization of out-of-the-box machine learning kits for specific purposes, such as micro-training a natural language processing system for analyzing financial contracts.
40:24Keith J. O'Donnellhighlights the importance of customizing machine learning models and modules for specific purposes and use cases.
40:43Keith J. O'Donnellemphasizes the need for technologists within the fintech industry to improve efficiency and optimization in their applications.
41:12Keith J. O'Donnellmentions the importance of computer vision and its industrial maturity for diverse use cases.
41:58Keith J. O'Donnelldiscusses the significance of natural language processing and its potential for increasing industrial collaboration.
42:13Keith J. O'Donnellinvites the community to contribute thoughts, ideas, and suggestions for the Zenith project.
42:39Nick Williamsraises a question about the security and veracity of data used in machine learning models.
43:10Keith J. O'Donnellacknowledges the importance of data security and welcomes feedback from the community.
43:14Keith J. O'Donnellsuggests making iterative updates to the primer based on community input and requirements.
45:00Keith J. O'Donnellprovides information on accessing the mailing list for updates and publications related to Zenith.
46:05Keith J. O'Donnellannounces the focus of the next deep dive session on commercialization and enterprise readiness.
46:44Keith J. O'Donnellencourages community members to contribute to the ongoing development and refinement of the primer.
47:25Keith J. O'Donnellopens the floor for any additional discussion or topics from the community.
51:32Velvárt Andrásinquires about getting started with the primer and suggests collaborative document creation.
52:41Keith J. O'Donnellacknowledges the need to get the primer document up and running and plans for further discussions on specific topics.
53:21Keith J. O'Donnellexpresses the goal of enhancing accessibility solutions through computer vision and collaboration.
53:59Keith J. O'Donnellinvites community input, ideas, and suggestions for the primer and Zenith project.
54:09Nick Williamsraises concerns about the security aspect of data used for financial advice and suggests addressing it in the primer.
54:36Keith J. O'Donnellacknowledges the importance of data security and encourages feedback and suggestions from the community.
56:19Keith J. O'Donnellemphasizes the iterative nature of the primer and welcomes community involvement and additions.