Open AI Seminars

Open AI Seminars

Photo

Welcome the Open AI Seminars, taking place every other Friday 13.00 to 14.00 at meeting room 454, IT-bygget, NTNU Gløshaugen

The seminars are open to everyone with an interest in AI and autonomous systems.

It is recommended to have some prior knowledge on AI.

The Open AI Seminars cover a wide range of topics within artificial intelligence, machine learning and autonomous systems. NTNU researchers, AI Lab partners, and invited speakers, will present their work and trigger discussions.

The objectives of these seminars are to bring together the research and innovation community at NTNU and beyond, who aims to apply AI in multidisciplinary research and industry, and spread creative ideas in order to challenge the entrenched notions of artificial intelligence. 

We will serve coffee and a sweet bite. 

We look forward to see you there!

Upcoming seminars

Upcoming seminars

Mini calevent portlet

Arrangement

Previous Seminars

Previous seminars

If you are interested in learning about previous seminars, you can find presentations from some of the past editions below: 

October 4: Open-Ended Machine Learning for Language Acquisition and Acoustic Reasoning.


October 4: Open-Ended Machine Learning for Language Acquisition and Acoustic Reasoning

In this talk, Prof. Giampiero Salvi (Dep of Electronic Systems, NTNU) gave an overview of his research aimed at modelling some aspects of early language acquisition and acoustic reasoning. The aim was not only to shed some light on how humans learn to interpret speech and acoustic events in general, but also to investigate problem settings that do not fit in the standard supervised machine learning framework.

If you are interested in learning more about this topic, you can take a look at his presentation below. 

April, 5: Machine Learning Methods in Market Research


April, 5: Machine Learning Methods in Market Research

In this seminar, Mike Kelly, Senior Group Director at Naxion Research & Consulting, addressed some of the business questions that market researchers commonly use machine learning models to help answer, such as: 

How can we best match our product/service offerings and associated messaging to the right customers? (machine learning application: Unsupervised learning with various types of clustering techniques). How can we accurately find customers (consumers or businesses) in attractive market segments? (machine learning application: discriminant analysis). What areas should we focus limited resources on to maximize improvement in customer satisfaction and retention? (machine learning application: random forests, Kruskal regression).

May 10


May 10, AI and blockchain in the energy sector

At this Open AI Seminar we explored how AI and blockchain technology can contribute to shape the energy sector of the future. Tárik Salem (PhD student at the Department of Computer Science) presented his paper and gave a talk on how to Forecast Imbalances in Electric Power Systems. In the second part of the seminar, Dr. Pedro Crespo Del Granado, gave a talk with the title: Blockchain based ElectricitY trading for the integration Of National and Decentralized local markets (BEYOND, EU project).

Open AI Seminar March, 29


March, 29: Unsupervised Learning in Manifolds

In his talk Prof. Zhirong Yang (NTNU), explored how we can identify the categories of data with little supervised information. Conventional cluster analysis approaches such as K-means, DBSCAN, spectral clustering or linkage methods do not work because real world data often distribute in curved manifolds. Prof.Yang recapped some of the previous work on how to recover the large-scale pattern, i.e., the clusters from massive local information, with a concrete example using MNIST handwritten digits. 

 

 

Artificial Intelligence in Theory and Practice


April, 12: Artificial Intelligence in Theory and Practice

In this talk, Professor and Director of the Norwegian Open AI Lab, Ole Jakob Mengshoel, presented his research on Stochastic local search (SLS) algorithms. These algorithms have proven to be very competitive in solving several computationally hard problems in artificial intelligence, machine learning, and signal processing. They perform well and can also be analyzed using Markov chains; this analysis brings key insight into their performance under varying conditions. The talk explored the foundations of SLS algorithms along with some applications.