Consider tasks like counting, sorting, and arithmetic. We humans learn algorithms for these processes so that we can apply them in all kinds of situations. For instance, while memorising the multiplication table works well for small numbers (0–9), learning the multiplication algorithm is far more efficient when dealing with larger numbers. This approach allows us … Continue reading Why is learning an algorithm sometimes better than learning from a large set of examples, as AI often does?
AI Letar Historiska Ansikten
Historikern Mats Hayen berättar om projektet Stadens Ansikten i den här SVT podden. Han berättat bland annat om hur AI hittar ansikten i olika foto-samlingar och de algoritmer jag har utvecklat i projektet.
Will AI eventually become more intelligent than Humans, as the Nobel laureate Geoffrey Hinton claims?
AI's performance is limited by the quality of the data it has been trained on—it can only be as good as its training data, never exceeding it. This limitation arises because AI excels at interpolation (making predictions within the range of its training data) but struggles with extrapolation (generalizing to entirely new or unknown situations). Additionally, AI lacks true … Continue reading Will AI eventually become more intelligent than Humans, as the Nobel laureate Geoffrey Hinton claims?
Visiting Lecturer
We had the pleasure of having Prof. Alex Telea from Utrecht University, visiting Uppsala 18-22 November and he gave several interesting lectures and talks. Among those was an InfraVis Colloquium where he presented 2D data visualisation techniques and also showed which high dimensional reduction techniques there are the most efficient. Interestingly, he showed that t-SNE … Continue reading Visiting Lecturer
Presentation at EB-CRIME
On the 29:th of October I presented my work on age estimation together with my colleagues in the EB-CRIME project. I use subspaces to improve performance and this is an ongoing research project.
Anders Hast 