#157 - AI in Theory and Practice - LØRN.TECH

Tema: AI

#157 – AI in Theory and Practice

Expørt: Ole Jakob Mengshoel

Professor

fra Dept. of Computer Science, NTNU

Med lørner Silvija Seres

Hva er de viktigste konseptene i AI? Og hvilke muligheter åpner denne teknologien opp for?  I denne episoden av #LØRN snakker Silvija og professor ved Dept. of Computer Science, NTNU, Ole Jakob Mengshoel, om gode eksempler på AI i teori og praksis.

Noen kjappe med ekspørt Ole Jakob Mengshoel

Ole Jakob Mengshoel  

XXX = AI 

Admin spørsmål (vi bruker disse for publisering): 

1) Navn, epost, mobil   [email protected]      

2) Selskap NTNU, CMU 

3) Tittel 

Professor, Dept. of Computer Science, NTNU 

Head, Norwegian Open AI Lab, NTNU 

Adjunct Faculty, Dept. of Electrical and Computer Engineering, CMU  

4) Ønsket overskrift 

Artificial Intelligence in Theory and Practice  

 

5) Tagger 

Artificial Intelligence, Machine Learning, Probabilistic Graphical Models, Deep Learning, Applications of Artificial Intelligence 

Podcast spørsmål (trenger også korte svar på disse): 

1) Hvem er du, og hvordan ble du interessert i XXX? 

My name is Ole Jakob Mengshoel, and I am a professor at NTNU.  

I became interested in artificial intelligence when I was an undergraduate student at NTH (now NTNU) in the 1980s.  

I remember well Tore Amble’s course on Logic Programming and a course on Expert Systems taught by researchers  

from SINTEF’s Knowledge-Based Systems group.  I then went on to take my Diploma Thesis with that group, and join them as an employee.  

2) Hva gjør dere på jobben? 

We work with a combination of theory, algorithms, software, and specific application areas (aerospace, medicine, telecom, …).  The specifics vary depending on the project; it may start with a more theoretical idea or a specific challenge tied to an interesting and important dataset.  

As a leader of the Norwegian Open AI Lab (see https://www.ntnu.edu/ailab) I often meet with interested parties inside and outside of NTNU to further our cause.  

3) Hva er de viktigste konseptene i XXX? 

Algorithms and data structures.  Programming. Probability theory and statistics.  

Mathematical maturity. An experimental mindset.   

4) Hvorfor er det spennende? 

I think artificial intelligence combines problems that are theoretically and computationally challenging with interesting and impactful applications.   The recent stronger focus on data-driven methods is also exciting.  

5) Hva synes du er de mest interessante kontroverser? 

As an example, the robot Sophia seems to receive a lot of press:  

https://en.wikipedia.org/wiki/Sophia_(robot) 

6) Dine egne prosjekter innen XXX? 

Among recent work (see https://works.bepress.com/ole_mengshoel/).  I will mention three projects:  

Improving safety for aerospace.  

Machine learning for human activity recognition.  

Integrating deep learning and multi-armed bandits for chatbots.  

7) Dine andre favoritteksempler på XXX internasjonalt og nasjonalt? 

There are many.  I will mention two:  

2005 DARPA grand challenge - five vehicles successfully completed a 212 km (132 mi) off-road course.  

Machine translation by Google - https://translate.google.com/ 

8) Hvordan pleier du å forklare XXX? 

Artificial Intelligence is focused on making smart computer systems.  

9) Hva gjør vi unikt godt i Norge av dette? 

I’ve just come back to Norway after 25 years in the US.  So I’m not in a good position to answer this question well.  

Generally speaking, Norway is focused on a comprehensive and inclusive process.  This has both positives and negatives.  

Products, projects, research, and services that leverage the positives and minimize the negatives will be at a potentially unique advantage.   

10) Anbefalt lesing/seing om XXX? 

It depends on your background and interests.  Here are a few examples.  

For the layperson:  

“The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World” Pedro Domingos 

For people with some relevant background:  

“Modeling and Reasoning with Bayesian Networks” Adnan Darwiche  

“Machine Learning: A Probabilistic Perspective” by Kevin P. Murphy 

11) En favoritt XXX sitat? N/A 

12) Viktigste poeng om XXX fra vår samtale? 

AI’s initial focus was on toy problems (e.g., the blocks world of Winograd and Winston), games, and puzzles.   

We can now (partly) solve much more complex AI problems as well, including autonomy, speech, and machine translation.  Forecasts of AI utopia or AI dystopia are currently far-fetched – let’s focus on using AI to solve society’s big problems, and improving the state-of- the-art in the process!  This is an  exciting time to engage with the Norwegian Open AI Lab, MS and PhD studies, and AI-related activities generally.  

 

 

Dette lørner du:

– Konseptene i AI
– Gode internasjonale og nasjonale eksempler på AI

Anbefalt litteratur:

Search name, company, subject or keyword