The document discusses applications of AI in various industries such as chemical, forest, pharmaceutical, and mining. It describes how AI techniques like machine learning, computer vision, and natural language processing can automate tasks and processes to increase efficiency. While AI brings benefits of greater efficiency and reducing risk, it also raises issues around responsibility and reducing human labor that require goal-oriented approaches and understanding of AI's impacts.
2. The “Magical” AI
- Digitalization, automation, data analytics, simulations, mathematical analysis
- Ongoing processes from 60’s
- Basic and current forms of AI (machine learning, computer vision, natural
language processing ) as new techniques and tools
- Breakthroughs: in machine learning, computation power, the amount of data
3. Some industry-related applications
- Machine learning, deep learning, neural networks
- Ability to learn without being explicitly programmed
- Classification, regression, clustering, anomaly detection
- Computer vision
- Automating what human visual system can do
- Natural Language Processing
- Reducing workload from researchers and employees
4. Reinforcement learning and other
advanced approaches
- A learning agent taking action towards maximizing the rewards
- An approach which full capacity we don’t know yet
- AlphaGo example: finding strategies humans haven’t ever found
- Experiments in video games: AI developing abilities not designed
- Highly funded research projects, but progressing
- Change of mindset: planning and designing goals and rewards
5. Business perspective
- Greater efficiency, reducing risk
- Tailoring, specializing, rapid changes
- As a tool for new R&D discoveries
- Releasing time for creativity (learning and studying, wellbeing at work, shorter
working hours?)
Societal and ethical perspectives
- The problem of responsibility (especially in medicine)
- Adding efficiency → Easiest way is reducing the human labor
6. 1. AI for industries
➔ Mostly good old technological
progress
➔ Take time to understand (data,
digitalization, AI) and follow
research.
➔ Goal-oriented approach in the
future
7. Resources
Nokia, Siilasmaa (so, don’t worry, you aren’t late)
https://blog.networks.nokia.com/innovation/2017/11/09/study-ai-machine-learning/
MOOC-courses (Coursera, Fast.ai)
https://www.coursera.org/learn/machine-learning
http://www.fast.ai/
Stanford - One Hundred Year Study on Artificial Intelligence (2016 report)
https://ai100.stanford.edu/2016-report
8. Collaboration in Turku region
Turku AI Society
- Connects researchers and students of AI impacts
aisociety.fi
Turku.ai Meetup
- Technical perspective (approx. monthly meetup with changing topics)
turku.ai
Research: TUGS, universities
Growing number of companies (some “AI” companies)