Open AI and Open Source AI (20240125@HKBU)

This presentation introduces Open AI and Open Source AI to MBA professors, alumni, and students, highlighting their differences and providing examples.

1. Open AI and Open Source AI,
what are the differences and examples?
Mr. Sammy Fung
Founder, Open Platform Society
President, Open Source Hong Kong
25 January 2024 in Hong Kong Bapist University

3. Sammy Fung
●Technology Consultant
–AI & ML
●Python
–Open Technology
●Open Data Research

4. Sammy Fung
●Open Source
–Hong Kong
●Founding Chairman, Hong Kong Linux User Group.
●President, Open Source Hong Kong.
–OSHK is an affiliate member of Open Source Initiative (OSI).
–Global
●Board Member (2022-2024), GNOME Foundation.
●Fellow Member (2022-), Python Software Foundation.
●Founder, Open Platform Society (2023-).

5. Sammy Fung
●Open Data
–Presentations:
●BarCamp HK 2013
●StartMeUp Week 2013 of InvestHK
●LegCo Secretariat Office
–Hackathons:
●Organiser
●Judge

6. Open
Source
Open AI
Open Content
Open
Hardware
Open
Governance
Open Data
Established in 2023
opsoc.org

7. Communications
●Emails:
–[email protected]
–[email protected]
●Websites:
–Open Platform Society: https://opsoc.org
–Open Source Hong Kong: https://opensource.hk

8. Let’s imagine about AI & our life
●Nowaday, “Life with AI” is irreversible.
●AI brings social benefits.
●How does AI change our life?
●How should we solve the problems with AI
technology in better way?

9. Generative AI
●
GenAI is Artificial Intelligence capable of generating text, images,
synthetic data, or other media, using generative models.
–
Machine Learning often uses statistical models.
●
Generative AI models learn the patterns and structure of their input
training data and then generate new data that has similar
characteristics.
●
Applications: writing, art, software development, etc.
●
LLM ChatBot Example: ChatGPT, Copilot, Bard, LlaMA
●
Text-to-Image Art System Example: Stable Diffusion, Midjourney,
DALL-E.

10. 7 Steps in AI Development Cycle
1.Defining objects and requirements.
2.Gathering Data.
3.Data Prepartaion and Manipulation.
4.Model Selection and Development.
5.Train the Data Model.
6.Validation and Testing.
7.Deployment and Maintenance.

11. What is Open AI?

13. What’s “Open AI”?
●San Francisco-based company OpenAI was created in 2015, its founders described
their mission:
What defines the ‘open’ in ‘open AI’? by Jennifer Ding, The Alan Turing Institute
https://www.turing.ac.uk/blog/what-defines-open-open-ai

14. Open AI
●
OpenAI was one of the first organisations to centre its
structure and public branding around the ‘open’
terminology.
●
In the years since, OpenAI has created revolutionary
algorithms, including the DALL·E and GPT model
families.
–
ChatGPT is built upon either GPT-3.5 or GPT-4, both of which
are members of OpenAI's proprietary series of generative pre-
trained transformer (GPT) models (from Wikipedia)
What defines the ‘open’ in ‘open AI’? by Jennifer Ding, The Alan Turing Institute
https://www.turing.ac.uk/blog/what-defines-open-open-ai

15. Open AI
●
However, along the way, the company has shifted from its original structure.
●
In 2019, it transformed from a non-profit to a “capped” for-profit
●
in 2020, gated its text-generating GPT-3 large language model (LLM) behind a
commercial API
–
soon after, it granted an exclusive license to GPT-3’s code to Microsoft, the company’s
biggest investor
●
These developments led some to question how open OpenAI’s practices really were,
reopening a conversation on what open means for AI development, beyond a value
signal to funders or great marketing copy.
●
OpenAI now seems to be taking DALL·E in a similarly commercial direction,
restricting free use and applying a ‘freemium’ business model.
What defines the ‘open’ in ‘open AI’? by Jennifer Ding, The Alan Turing Institute
https://www.turing.ac.uk/blog/what-defines-open-open-ai

16. `
https://www.technologyreview.com/2020/09/23/1008729/openai-is-giving-microsoft-exclusive-access-to-its-gpt-3-language-model/

17. What does this “exclusive” mean?
●The companies say OpenAI will continue to
offer its public-facing API, which allows chosen
users to send text to GPT-3 or OpenAI’s other
models and receive its output. Only Microsoft,
however, will have access to GPT-3’s
underlying code, allowing it to embed,
repurpose, and modify the model as it pleases.
https://www.technologyreview.com/2020/09/23/1008729/openai-is-giving-microsoft-exclusive-access-to-its-gpt-3-language-model/

20. GDPR and AI
●The European Union's General Data Protection
Regulation (GDPR) is considering AI
regulations. GDPR's strict limits on how
enterprises can use consumer data already
limits the training and functionality of many
consumer-facing AI applications.

21. Code is Law
●
Lawrence Lessig
–
Professor of Law at Harvard Law School.
–
Founder of Creative Commons – open licenses for digital contents.
●
A Book: Code and Other Laws of Cyberspace
●
a form of regulation where private actors may embed their
values into technological artifacts, effectively constraining our
actions.
●
AI ? Blockchain (aka Web3) ?

23. UK Post Office Scandal
●Between 1999 and 2015, more than 900 sub-
postmasters and postmistresses were prosecuted
for theft and false accounting after money appeared
to be missing from their branches, but the
prosecutions were based on evidence from faulty
Horizon software.
●Some sub-postmasters wrongfully went to prison, many
were financially ruined. Some have since died.

24. UK Post Office Scandal
●
Fujitsu Europe's boss has admitted the firm has a "moral
obligation" to contribute to compensation for sub-postmasters
wrongly prosecuted as a result of its faulty IT software.
●
Paul Patterson said Fujitsu gave evidence to the Post Office
that was used to prosecute innocent managers.
●
He added that the Post Office knew about "bugs and errors" in
its Horizon accountancy software early on.
●
The global chief executive of Fujitsu, Takahito Tokita, also
apologised.

27. AI Alliance
●Facebook parent Meta and IBM on Tuesday launched a new group
called the AI Alliance that’s advocating for an “open science”
approach to AI development that puts them at odds with rivals
Google, Microsoft and ChatGPT-maker OpenAI.
●The AI Alliance — led by IBM and Meta and including Dell, Sony,
chipmakers AMD and Intel and several universities and AI startups —
is “coming together to articulate, simply put, that the future of AI is
going to be built fundamentally on top of the open scientific
exchange of ideas and on open innovation, including open
source and open technologies,”

29. Is Llama 2 open ?
●
In July, Meta released its large language model Llama 2 relatively openly and
for free, a stark contrast to its biggest competitors. But in the world of open-
source software, some still see the company’s openness with an asterisk.
●
While Meta’s license makes Llama 2 free for many, it’s still a limited license
that doesn’t meet all the requirements of the Open Source Initiative (OSI)
●
Meta’s limits include requiring a license fee for any developers with more
than 700 million daily users and disallowing other models from training on
Llama.
●
IEEE Spectrum wrote researchers from Radboud University in the Netherlands
claimed Meta saying Llama 2 is open-source “is misleading,” and social media
posts questioned how Meta could claim it as open-source.
https://www.theverge.com/2023/10/30/23935587/meta-generative-ai-models-open-source

30. Open Source Definations &
Licenses
●Some keywords of Open Source Definations
–Source Code: Free Redistribution, Allow Derived Works
–Distribution of (Open Source) License
●Open Source Initiative (OSI) approves open source
licenses that comply with the Open Source Definition.
–Some of Popular Licenses
●GNU GPL, BSD, Apache, MIT Licenses

32. Is Falcon LLM open ?
●
Developed by Technology Innovation Institute (TII), United Arab of Emirates.
●
Falcon 180B is an open-access large language model that builds on the previous releases
in the “Falcon” family. It's a scaled-up version of the Falcon 40B model, an AI solution that
ascended to the top of the Hugging Face LLM Leaderboard in May 2023.
●
Falcon 40B is made available under the permissive Apache 2.0 software license for
researchers and commercial users alike.
●
Falcon-180B is accessible to developers through a royalty-free license, based on
Apache 2.0.
●
Those hosting providers wishing to provide shared instances of the model or its
derivatives as a managed service (for inference or fine tuning) are not covered by
the proposed license, and are invited to enter into a separate license arrangement with
TII.

33. Open Source AI Tools
to train AI models
●TensorFlow
●PyTorch
●Keras
●Scikit-learn
●OpenCV

34. TensorFlow
●
TensorFlow allows programmers to construct and
deploy machine learning models across various
platforms and devices.
●
Its robust community support and extensive library of
pre-built models and tools streamline the development
proces
●
making it easier for beginners and experienced
practitioners to innovate and experiment with AI.

35. PyTorch
●
PyTorch is an open-source AI framework offering an intuitive
interface that enables easier debugging and a more flexible
approach to building deep learning models.
●
Its strong integration with Python libraries and support for GPU
acceleration ensures efficient model training and
experimentation.
●
It is a popular choice among researchers and developers for
rapid software development prototyping and AI and deep
learning research.

36. Keras
●
open-source neural network library written in Python
●
It’s known for its user-friendliness and modularity,
allowing for easy and fast prototyping of deep learning
models.
●
It stands out for its high-level API, which is intuitive for
beginners while remaining flexible and powerful for
advanced users, making it a popular choice for
educational purposes and complex deep-learning
tasks.

37. Scikit-learn
●
a powerful open-source Python library for machine learning and
predictive data analysis.
●
Providing scalable supervised and unsupervised learning
algorithms,
●
it has been instrumental in the AI systems of major companies
like J.P. Morgan and Spotify.
●
Its simple setup, reusable components and large, active
community make it accessible and efficient for data mining and
analysis across various contexts.

38. OpenCV
●
a library of programming functions with comprehensive
computer vision capabilities, real-time performance,
large community and platform compatibility
●
an ideal choice for organizations seeking to automate
tasks, analyze visual data and build innovative
solutions.
●
Its scalability allows it to grow with organizational
needs, making it suitable for startups and large
enterprises.

40. Hugging Face
●French-American company based in New York
City.
●The AI community platform contains Models
and Datasets.
–“GitHub” of AI models using Git version control.
–Datasets are mainly in text, images, and audio.

42. Deep Dive

43. Deep Dive:
Process to defining Open Source AI
●Open Process: This Deep Dive is a series of events, in person
and online, gathering opinions and comments from individual
developers, lawyers, researchers, non-profit organizations,
companies and government officials developing and using AI
systems.
●OSI is bringing together global experts to establish a shared
set of principles that can recreate the permissionless,
pragmatic and simplified collaboration for AI practitioners,
similar to what the Open Source Definition has done.
https://opensource.org/deepdive/

44. Deep Dive Timeline

45. Deep Dive Timeline

46. The Open Source AI Definition
●Draft version 0.0.4.
●the definition of AI system adopted by the Organization for
Economic and Co-operation Development (OECD)
–An AI system is a machine-based system that, for explicit or
implicit objectives, infers, from the input it receives, how to
generate outputs such as predictions, content, recommendations,
or decisions that can influence physical or virtual environments.
Different AI systems vary in their levels of autonomy and
adaptiveness after deployment.
https://blog.opensource.org/closing-the-2023-rounds-of-deep-dive-ai-with-first-draft-piece-of-the-definition-of-open-source-ai/

47. What is Open Source AI ?
●
The Open Source AI Definition - Draft version 0.0.4.
●
To be Open Source, an AI system needs to be available under
legal terms that grant the freedoms to:
–
Use the system for any purpose and without having to ask for
permission.
–
Study how the system works and inspect its components.
–
Modify the system for any purpose, including to change its output
–
Share the system for others to use with or without modifications, for
any purpose.
https://blog.opensource.org/closing-the-2023-rounds-of-deep-dive-ai-with-first-draft-piece-of-the-definition-of-open-source-ai/

48. Status & Plan of
The Open Source AI Definition
●
Draft version 0.0.4.
●
Still in review, writing other parts.
●
working on a plan for 2024 that includes
expanding our reach to other communities with
an eye on reaching consensus on a 1.0 release
of the Open Source AI Definition in the quickest
amount of time.
https://blog.opensource.org/closing-the-2023-rounds-of-deep-dive-ai-with-first-draft-piece-of-the-definition-of-open-source-ai/

49. Recaps
●AI development process.
●What’s Open AI ? And is Open AI “open” ?
●Should we totally trust AI ? and examples.
●Some other “Open” AI models.
●Some Open Source AI Tools.
●Open Source AI Definition.

50. At the end, let’s imagine again
●Nowaday, “Life with AI” is irreversible.
●AI brings social benefits.
●How does AI change our life?
–And any new challenges today with AI ?
●How should we solve the problems with AI technology in
better way?
–Open Source AI ?