Ai ml dl_bct and mariners-1
- 1. AI/ML/DL/BCT A Revolution in Maritime Sector By: Capt. Krishnamurthy Iyer Dr. Sanjeet Kanungo Mr. Ganesh Ingle
- 2. Session objective An introduction to AI,ML,DL Working of AI System Scope of AI ,Cyber Security and BCT in Marine Education Changes Required in Curriculum Summary
- 3. New Advancement in Marine Field 4 Technologies will revolutionize the maritime sector 1. AI (Artificial Intelligence, Machine learning and Deep learning) 2. Industrial IOT 3. Block Chain Technology(BCT) 4. Virtual Reality 5. Cyber Threat and Cyber Security
- 4. Scope of AI and BCT 1. Supply chain visibility/ Logistic management 2. ML-driven data science 3. Voyage optimization 4. Voyage economics 5. Diagnosis systems - infer malfunction of devices and components 6. Monitoring - compare data from a continually observed system to prescribe behaviour 7. Process control - control a physical process based on monitoring 8. Design - configure a system according to specifications 9. Scheduling & planning - develop or modify a plan of action 10. Generation of options - generate alternative solutions to a problem New Advancement in Marine Field
- 5. AI based Systems 1. Google Search Engine 2. Microsoft Cortana 3. Amazon Alexa 4. Image Processing 5. Autonomous cars : Tesla( Same AI based system will be implemented in navigation AI based system)
- 6. AI 1. Do you see cat or a dog??? 2. How did you analyse? 3. As you have seen many cats and dogs 4. You learnt to identify the difference between them i.e. pattern recognition 5. This how we train the computer to recognize pattern and learn 7. Sometime human could not recognise the pattern 8. Same thing can happen with AI
- 7. How Do Our Brains Work? Dendrites: Input Cell body: Processor Synaptic: Link Axon: Output Can we create artificial neurons??? Image Source: Wikiped
- 8. How Do ANNs Work? Inputs Weights Bias value Hidden layer Sigmoid function Activation function Back propagation Output Fitness function
- 9. Artificial Intelligence AI Act Like Human Turing Test Playing chess by machine and human in different rooms with observer not knowing which room has machine Complete Turring Test NLP: Natural language representation KR: Knowledge representation Movement: Robotics Rationally Rational Agent approach Think Like Human Cognitive Approach ANN design ANN training Activation functions Weights to the ANN Rationally Law of Thought Approach i/p 1=Ram is human i/p =All human are kind o/p=Ram is Kind
- 10. AI Hard AI Soft AI Artificial Intelligence Machine learning algorithms are the latest in computer science where computers can take a large data set, analyse it and use it to make increasingly accurate predictions. In short, they are learning. To create a self-driving car : 1. Hundreds of thousands, or even millions, of miles of driving videos and data input to data model. 2. It creates a massive vocabulary around driving. 3. The algorithms use visual techniques to break down the videos and to understand them. 4. The goal is that when something unexpected happens — a ball rolls into the street — the car can recognize the pattern and react accordingly (slow down because a child could be running into the street after it). 5. Same algorithms is applied for autonomous ships
- 11. ML Feature extraction and compare
- 12. DL No feature extraction required system recognizes pattern on its own Complex system of algorithm
- 13. DL Complex system of algorithm
- 14. Steps Involved in AI ,ML and DL 1) Determine your inputs: 2) Assign random weights: 3) Train the data: 4) Utilize the model: Image Source: Wikipedia
- 15. AI ,ML and Deep learning Image Source:www.discuss.analyticsvidya.com
- 16. ML and Deep learning
- 17. Learning by Trial‐and‐Error Continuous process of: Trial: Processing an input to produce an output (In terms of ANN: Compute the output function of a given input) Evaluate: Evaluating this output by comparing the actual output with the expected output. Adjust: Adjust the weights.
- 18. 1. Supervised Training. Inputs and the outputs are provided by user Compares resulting outputs w.r.t. desired outputs Errors are back propagated, causing the system to adjust the weights which control the network. 2. Unsupervised, or Adaptive Training. Inputs dataset but no desired outputs for comparison System itself must decide what features will use to group the input data Present time, unsupervised learning is not well understood ANN Training
- 19. CHARACTERISTICS TRADITIONAL COMPUTING (including Expert Systems) ARTIFICIAL NEURAL NETWORKS Processing style Functions Sequential Logically (left brained) via Rules Concepts Calculations Parallel Gestault (right brained) via Images Pictures Controls Learning Method Applications by rules (didactically) Accounting word processing math inventory digital communications by example (Socratically) Sensor processing speech recognition pattern recognition text recognition ANN Training
- 20. AI in Merchant Navy 1. Are the ships completely AI based now??? 1. The ships will become fully autonomous when the large set of data is provided to the ML and DL system. 2. After processing and training (unsupervised) is done, fully autonomous ships will sail without manual interventions. 3. Till the system get fully autonomous till the training/ learning part of ML/DL human interventions is required to for manual override(Manual override can be done from shore based control centre.
- 21. AI in Merchant Navy 2. When the ships will be completely AI based?? 1. Once the ML, DL based algorithms are fully trained on the large scale database with almost all environmental conditions then the system ships will be totally autonomous. 2. Till the time manual override is must
- 22. Cyber Security 1. Cyber world is world of binary no physical lock does exists. 2. Google, Amazon, Microsoft and Facebook. Estimates are that the big four store at least 1,200 petabytes between them. That is 1.2 million terabytes (one terabyte is 1,000 gigabytes). 3. If the systems including AI based are infected by virus or attacked by hackers then complete system will collapse.
- 23. DOS DDOS Brut force Phishing attack SQL injection attack Cross site scripting attack Eavesdropping attack. Password attack Birthday attack Malware attack Replay attack Different Types of Cyber Attacks Mounting a good defence requires understanding of the offense. Caller ID/Email/SMS spoofing attack Man in middle attack (Session hijacking ) IP spoofing attack (Replay attack)
- 24. A shipping company was attacked by Ransom ware Phishing mail attack Computers were infected that aren’t patched, and then passwords hacked from those computers to infect other computers that are patched. Exploited complete active directory. Exploited DC. Unconditional jump from n/w to n/w Encrypted MBR Systems infected >1 million domain controller @ Ghana Mirroring isolation backup Company required reinstall 4000 servers, 45000 PCs Ransomware attack would cause losses of up to $300 million A month after cyber attack launched named wannacry1,2 A Case Study of Recent Cyber-Attack NotPetya EternalBlue Vulnerability in windows/any OS Srv.sys,445 port Mimikatz French security researcher Benjamin(Password)
- 25. 1. Prerequisites required for AI based System i. Maths Domain: Mathematical logic/ Propositional Logic-PL, Predicated logic / First order predicate logic (FOL), SET theory,, Resolution, FOL, Learn linear algebra, probability, multivariate calculus, optimization and few other topics ii. Computer domain : Data structures, programming language like C,C++,Python, Go (any one language).Cyber threat and cyber security awareness. 2. Seafarer: Undergraduates (10+2 pass) must know after graduation/DNS i. Existing syllabus with following points are required ii. Over view and working of AI based system and automation. iii. AI based system simulators training. iv. Skill based maintenance and error code analysis generated by AI based system on board ship. v. Manual override in during the AI/ML/DL system training phase vi. IOT , computer hardware familiarization and maintenance. vii. Maintenance knowledge of PLC,SCADA,DCS systems. viii. Hydraulic and pneumatic control systems governed by AI Prerequisites of AI for Seafarer
- 26. Prerequisites required for AI based System i. Maths Domain: Mathematical logic/ Propositional Logic-PL, Predicated logic / First order predicate logic (FOL), SET theory,, Resolution, FOL, Learn linear algebra, probability, multivariate calculus, optimization and few other topics ii. Computer domain : Data structures, programming language like C,C++,Python, Go (any one language).Cyber threat and cyber security awareness. 1. Controlling from shore :Entry Level Undergraduates (10+2 pass) i. Existing syllabus with following points are required ii. AI based system simulators training iii. Manual override and system maintenance and upgrade iv. Skill based maintenance and error code analysis generated by AI based system. v. IOT , computer hardware familiarization and maintenance. vi. Maintenance knowledge of PLC,SCADA,DCS systems vii. Training the AI based system and data acquisition from thousands of data node and updating the system experience. Prerequisites of AI for Seafarer
- 27. Fundamentals Remain Same To operate a machinery or equipment the basis of training to remain same. Prerequisites of AI for Seafarer
- 28. Manual overrides The shipboard personnel to be trained additionally on taking over the controls when required. The task is not simple as varied inputs are being fed for a single equipment or machinery Automation can be handled but Artificial operation needs critical thinking. Should the AI system give erroneous output, the shipboard personnel should be able to detect the deviation. Prerequisites of AI for Seafarer
- 29. Parametric Analysis The real time parameters to be analysed at defined intervals to prove the system efficacy. This brings one clear add on in the syllabus. Extensive use of MATLAB can help understand the neural network to understand and thus compare the inputs versus output. The troubleshooting currently in the training methodology to be further augmented to numeric comparison to help the students correlate the inputs vs output and thus identify the degree of severity. Prerequisites of AI for Seafarer
- 30. Skill Set Requirement • As said the maximum accidents are due to HUMAN ERROR, the concept of AI operated vessels might give some disastrous results if not understood well. • It is to be borne in mind that the neural network inputs are based on the ecosystem the machine or equipment is operating and behavioural patterns will be different even for the same machine. Prerequisites of AI for Seafarer
- 31. Response • AI researchers want to study AI the same way social scientists study humans. • Propose a new academic semester called ''Machine Behaviour'', which will be through empirical observations and experimentation. • The rise of machines with agency, i.e. machines that are actors making decisions and taking actions autonomously. Thus they need to be studied '' as a new class of actors with their own behavioural patterns and ecology'' Prerequisites of AI for Seafarer
- 32. Response • This does not mean to suggest that AI systems have developed some kind of free will. • They certainly have not; they are only glorified math model. • But it is meant to move away from viewing AI systems as passive tolls that can be assessed purely through their technical architecture, performance and capabilities. • They should be instead be considered as active actors that change and influence their environments and the people and machines around them. Prerequisites of AI for Seafarer
- 33. Response • The study to aim at emergent properties that arise from many humans and machines coexisting and collaborating together. • '' Figuring out how to live with machines is a problem too vast for any one discipline to solve alone''----- So, bringing them together under one umbrella , helps take advantage of the other's expertise , the more they will be able to ensure that, the artificial agents benefit humans rather than harm them. Prerequisites of AI for Seafarer
- 34. Types of Jobs There will be two types of jobs • Operators • Programmers and shore control • AI should more correctly be described as Assisted Intelligence – Robotic systems will be controlled and managed by intelligent humans till the system gets fully trained.
- 35. Change Required in Syllabus • In BSc Syllabus increase mathematics portion and electronics portion, hardware maintenance, cyber security awareness • Should be able to understand circuit diagram • Control system should also be included in the syllabus • Basic fundamental subjects should be included like mechanics, hydraulics Existing officers have a module to cover Maths, Electronics, Computer software and hardware
- 36. Safe Return to Port • Vessel should be designed taking into account Safe Return to Port which is now applicable to Passenger vessels. • Full redundancy for propulsion and electrical production will be required as well as for Steering system
- 37. Safe Return to Port • In case of casualty affecting the bridge, an alternative place shall be arranged where essential equipment (fixed or portable) for navigation. • We recommend for ECDIS the software and the chart should be loaded in a laptop and kept in master’s cabin.
- 38. Summary • AI is the future of the marine industry or any other industry • The pace of progress of AI is incredibly fast • We have accept the fact and upgrade accordingly • It is going to be interesting to see how society deals with AI but definitely it will be cool