Artificial Intelligence- Introduction and Applications.pdf
- 1. Artificial Intelligence Anand J Kulkarni PhD, MS, BEng, DME Research Professor & Associate Director Institute of Artificial Intelligence Dr Vishwanath Karad MIT World Peace University Pune 411 038, MH, Bharat (India) Email: anand.j.kulkarni@mitwpu.edu.in; kulk0003@ntu.edu.sg Ph: 91 70301 29900 Dr Anand J Kulkarni 1
- 2. Speaker Information Post Doctorate Research Fellow (Cross Border Supply Chain Management Project) PhD (in Nature Inspired Optimization) MASc (in Artificial Intelligence) BEng (in Mechanical Eng) Diploma (in Mechanical Eng) Dr Anand J Kulkarni 2
- 3. Speaker Information AI optimization methodologies/Theories Pioneered 1. Probability Collectives 2. Cohort Intelligence 3. Expectation Algorithm 4. Socio Evolution & Learning Optimization Algorithm 5. Ideology Algorithm 6. LAB Algorithm 7. Golf Sport Inspired Search Algorithm Journal Articles: 36 Book Chapters: 25 Conference Papers: 30 News Paper Articles: 35 3 Dr Anand J Kulkarni
- 4. Speaker Information Editorial Activities Lead Editor: Engineering Optimization: Methods and Applications Book Series Editors: Kulkarni, A.J., Gandomi, A.H., Mirjalili, S.A., Lagaros, N.D., Liao, T.W. Advances in Metaheuristics Book Series Editors: Kulkarni, A.J. & Siarry, P. Associate Editor Engineering Applications of Artificial Intelligence (Elsevier) Systems and Soft Computing (Elsevier) 4
- 5. Speaker Information Advances in Metaheuristics Book Series Editors: Kulkarni, A.J. & Siarry, P. Editorial Activities Engineering Optimization: Methods and Applications Book Series Editors: Kulkarni, A.J., Gandomi, A.H., Mirjalili, S.A., Lagaros, N.D., Liao, T.W. 5
- 6. Dr Anand J Kulkarni 6
- 7. Data Science-AI-ML-Statistics Dr Anand J Kulkarni 7
- 8. What is Intelligence • Every living creature’s intelligence is characterized by • Perception • Reasoning • Inference • Learn • Imitate/Mimic/Follow • Creative • Adapt Dr Anand J Kulkarni 8
- 9. Intelligence: Perception Dr Anand J Kulkarni 9
- 10. Intelligence: Perception Dr Anand J Kulkarni 10
- 11. Intelligence: Reasoning Dr Anand J Kulkarni 11
- 12. Dr Anand J Kulkarni 12
- 13. Intelligence: Inference • Making inferences is like being a detective; you may look at all the details and put them together to draw a meaningful conclusion. Dr Anand J Kulkarni 13
- 14. Intelligence: Mimic/Copy/Follow/Imitate Dr Anand J Kulkarni 14
- 15. Intelligence: Adapt and Evolve Dr Anand J Kulkarni 15
- 16. Intelligence: Adapt and Evolve Dr Anand J Kulkarni 16
- 17. Dr Anand J Kulkarni 17
- 18. Darwin’s Theory of Adaptation and Evolution 1809-1882 Dr Anand J Kulkarni 18
- 19. History: Natural Intelligence into Artificial Intelligence (AI) • Dr Alan Turing (1950): Computing Machinery and Intelligence • Very High computer prices (in several million rupees) • Computer with no Memory • Concept of (humanlike) Thinking Machines initialized • Scope limited to Robots and Automation Dr Anand J Kulkarni 19
- 20. History: Natural Intelligence into Artificial Intelligence (AI) • Dr John McCarthy (1956): Dartmouth Summer Project on AI • Invited 11 renowned mathematicians, professors and economists Dr Marvin Minsky Dr Julian Bigelow Dr D.M. Mackay Dr Ray Solomonoff Dr John Holland Dr John McCarthy Dr Claude Shannon Dr Nathanial Rochester Dr Oliver Selfridge Dr Allen Newell Dr Herbert Simon Dr Anand J Kulkarni 20
- 21. Fifth Generation Project Dr Anand J Kulkarni 21
- 22. Intelligence Everywhere Dr Anand J Kulkarni 22
- 23. Anand J Kulkarni, PhD 23 1952 1957 1962 1967 1971 1977 1980 1984 1989 1991 1996 1998 1999 2004 2009 2014 2019 Ideology Algorithm 21-12-2023
- 24. Artificial Intelligence Everywhere Dr Anand J Kulkarni 24
- 25. • Machine Learning for recognition - analyzing animal photos, behavior video, pugmarks - Cogito Tech Drones • Marc Travers (Avian Conservation) using Conservation Metrics • LoRaWAN Dr Anand J Kulkarni 25 Rainforest Connection, Sensing Clues, Open Ears, Audio Moth for animal, bullet firing acoustics classification Uof Southern California for Wild Life Conservation in Cambodia
- 26. Dr Anand J Kulkarni 26 Durham police and Judiciary System for Parole decision
- 27. Dr Anand J Kulkarni 27 2017 – Turnover on Smart Homes: INR 150 Billion 2024 – Expected > INR 1000 Billion
- 28. Dr Anand J Kulkarni 28 Everyday waste > 200 billion tons Zen Robotics, Finland Average human: 30-40 objects Robots: 150-160 AMP Robotics, USA classifying certain brand packings Bin-E, Poland Wastebox, Austria
- 29. Dr Anand J Kulkarni 29 Population today: 7 Billion 2050: 9 Billion Mordor Intelligence, UK: AI Investment in Food Business 2020: INR 20 Billion 2026: INR 2100 Billion TOMRA, Norway Winnow Solutions, UK Kitchen Cameras, Shanghai Emaar Hospitality, Dubai Wastage reduced by 70%
- 30. Dr Anand J Kulkarni 30 WARDSAUTO Newsletter Today: 1.2 Billion Vehicles 2035: 2 Billion Vehicles Potential Gridlock
- 31. Dr Anand J Kulkarni 31 Artificial Intelligence Everywhere
- 32. AI in/for Education: Connecting the Dots Dr Anand J Kulkarni 32
- 33. AI Driven Cafeteria Approach • Student’s Choice/Interest at the centre Dr Anand J Kulkarni 33 Local & Global Universities & Institutes Industries University/Institute Courses Dynamics and vibration Engineering design Heat and mass transfer Fluid mechanics Systems, measurements, and controls Solid mechanics … Thermodynamics Industry Courses Power Train maintenance Switch Gear Design … Automobile Engineering
- 34. Dr Anand J Kulkarni 34
- 35. Why AI? Dr Anand J Kulkarni 35 • Healthcare • The problem arises in minimizing congestion in the recovery unit • Planning horizon of n days with cyclic scheduling • Keep the maximum number of patients as low as possible to reduce the requirement of beds, nurses and other variable costs 40 30 22 13 36 29 23 7 32 30 22 7 36 32 23 17 𝑴 𝑻 𝑾 𝑹 𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 𝐼𝑘 = 144 121 90 44 𝑍 𝐶∗𝜋1 = 105 22 13 40 30 29 23 7 36 7 32 30 22 36 32 23 17 𝑴 𝑻 𝑾 𝑹 𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 𝐼𝑘 = 94 100 100 105 𝜋∗1 = 3, 4, 2, 1 Beds, Nurses, Variable Costs, etc.
- 36. Dr Anand J Kulkarni 36 Problem Size Average % gap LP vs CPLEX CPLEX: Average CPU Time (Sec) CI Method Average % gap LP vs CI Average % gap CPLEX vs CI Standard Deviation (SD) CPU Time (Sec) Function Evaluations (FE) 5 x 5 4.3165 0.27 4.5496 0 0 0.038084 1620 6 x 6 3.3520 0.29 3.4857 0 0 0.082466 3169 7 x 7 2.4426 0.41 2.5067 0 0 0.141646 5070 8 x 8 1.7478 0.73 1.7815 0 0.659745 0.206037 6988 9 x 9 1.3584 4.10 1.3778 0 1.920292 0.301112 9136 10 x 10 0.9259 13.80 0.9473 0.0120 3.265155 0.359756 10750 11 x 11 0.8597 98.31 0.9739 0.1057 2.961605 0.443356 12190 12 x 12 0.7083 260.59 0.9296 0.2146 3.045473 0.537049 13601 13 x 13 0.5632 1072.93 0.8675 0.2993 2.695255 0.663312 15477 15 x 15 -- -- 0.8040 -- 2.832436 0.831481 17827 20 x 20 -- -- 0.6492 -- 2.727449 1.659961 26085 25 x 25 -- -- 0.5431 -- 2.636055 3.138706 35435 30 x 30 -- -- 0.4502 -- 2.668305 5.005378 43715 35 x 35 -- -- 0.4031 -- 2.586950 7.842689 51145 40 x 40 -- -- 0.3417 -- 2.696257 11.972532 61610 45 x 45 -- -- 0.3130 -- 2.468574 17.019936 71835 50 x 50 -- -- 0.2784 -- 2.545943 23.980858 81207
- 37. Generative AI • Self Supervised learning • LLM: Trained on Heavy data (Google, Meta, etc.) • Image Based • New Content Generation • Images • Audio/Music • Video • Text • Speeches Dr Anand J Kulkarni 37
- 38. Dr Anand J Kulkarni 38 https://www.nytimes.com/2022/09/02/technology/ai-artificial-intelligence-artists.html Jason M. Allen via Midjourney Jason Allen’s A.I.-generated work, “Théâtre D’opéra Spatial,” took first place in the digital category at the Colorado State Fair 2022
- 39. OpenAI: DALL E • Creates original, realistic images and art from a text description by combining concepts, attributes and styles. • DALLE E employs Two Techniques • CLIP: Learns relationship between images and associated text and makes the computer learn the concepts in images and hence can generate new images of the same concepts Dr Anand J Kulkarni 39
- 40. • CLIP • Diffusion An Astronaut Riding a Horse Added Gaussian Noise Noise Cleared Dr Anand J Kulkarni 40
- 41. Dr Anand J Kulkarni 41
- 42. Dr Anand J Kulkarni 42
- 43. The transformation is determined by the captions “a Victorian House”, which describes the architecture of the house, and “a Modern House”, which describes how the architecture of the house should be changed. Dr Anand J Kulkarni 43
- 44. Dr Anand J Kulkarni 44
- 45. GAN Dr Anand J Kulkarni 45
- 46. chAIr Dr Anand J Kulkarni 46 https://steffen-weiss.design/the-chair-project-generating-a-classic GAN was used with a dataset of iconic 20th- century chairs with the goal to `generate a classic`. The resulting model was used to generate new chairs as visual prompts for a human designer who turned these images into feasible chairs through manual sketching and rendering. chAIr project explores co-creativity between humans (Philipp Schmitt, Steffen Weiss) and AI (ANN), taking the chair — the archetype/illustrative model of a designed object — as an example.
- 47. Fights Counterfeiting Dr Anand J Kulkarni 47
- 48. Competency Dr Anand J Kulkarni 48 https://openai.com/research/gpt-4
- 49. Competency Dr Anand J Kulkarni 49 https://openai.com/research/gpt-4
- 50. Training & Assistance • Q&A by the E-Trainer or Expert Trainer Dr Anand J Kulkarni 50
- 51. Training & Assistance Dr Anand J Kulkarni 51 𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑓 𝑥 = 𝑥2 , −2 ≤ 𝑥 ≤ 2
- 52. Generating Reports/Manuscripts • Trained on the data till 2021 (*) • Cannot pull the references (*) • Could be erroneous (*) • Generate Literature Review • (Text length improved from 3000 to 25000 words) • Over dependency on AI generating Emails, Reports, Manuscripts, etc. Dr Anand J Kulkarni 52
- 53. Generating Class Work • Generate assignments • Generate assignment Solutions • Generate computer codes • Must stay at least few steps ahead in understanding and acquiring the knowledge. Dr Anand J Kulkarni 53
- 54. Challenges • Deep Fake • Polarization of the communities 54 21-12-2023 Anand J Kulkarni, PhD
- 55. Challenges and Solutions • Explainable AI • Ethical AI • Teaching & Leaning Dr Anand J Kulkarni 55
- 56. AI Contributing: Go Green and Sustainability • According to UN Alliance for Sustainable Fashion Apparel manufacturing accounts for • 8% of the world’s greenhouse gas emissions • 9% of annual microplastics in the oceans • Restocking returns is often not financially feasible and most end up in a landfill • Traditional approach: Deciding what customers want, then mass producing that item, is a game of chance: Bet wrong, and a manufacturer generally ends up with a lot of unsold clothing. • AI can help on several fronts • Starting with trends forecasting • Several firms use AI and ML to analyze images on social media, taking note of prints, shapes and color to help their manufacturer clients figure out what’s going to sail and what’s going to sink • AI helps the brands figure out pricing strategies and steer clear of trends that are on their way out • Fewer returns • Reduced waste • Increasing industry’s sustainability Dr Anand J Kulkarni 56
- 57. Thank you Dr Anand J Kulkarni 57 Anand J Kulkarni PhD, MS, BEng, DME Research Professor & Associate Director Institute of Artificial Intelligence AI 502 Avanti, MIT World Peace University 124 Paud Road, Kothrud, Pune 411038, MH, Bharat (India) Email(s): anandmasc@gmail.com; anand.j.kulkarni@mitwpu.edu.in; kulk0003@ntu.edu.sg URL: sites.google.com/site/oatresearch/anand-jayant-kulkarni