About

Intelligent automated systems are rapidly penetrating all areas of our life. To manage increasingly complex interactions with humans around, machines and algorithms often need to be able to anticipate the behavior of these humans. At AiTech, I explore ways to enable artificial agents to anticipate humans’ decisions using cognitive models of decision making, building up on the extensive body of knowledge elaborated in cognitive psychology and neuroscience. I believe that insights from these fields can help us to develop, for instance, autonomous vehicles which would reliably predict the decisions of human agents during traffic interactions. The overarching goal of my work is to keep the complex interactions between humans and artificial agents under meaningful human control.

E-mail: a.zgonnikov at tudelft dot nl

google scholar | open science framework | github

twitter | orcid | researchgate

Arkady ZgonnikovFollow

Researcher @aitech_tudelft @tudelft. Decision making, driver behavior, meaningful human control over autonomous systems. And some french fries, please.

Retweet on Twitter Arkady Zgonnikov Retweeted

Gustav Markkula @markkula·

18h

A central realisation is that regardless of whether one's perspective is from game theory, traffic safety, sociology, linguistics, ... traffic interactions are fundamentally about what we term "space-sharing conflicts" - we identify some prototypes:

Reply on Twitter 1221542589305847811 Retweet on Twitter 12215425893058478111 Like on Twitter 12215425893058478112 Twitter 1221542589305847811

Retweet on Twitter Arkady Zgonnikov Retweeted

Gustav Markkula @markkula·

18h

Preprint! Bc of AVs, research on interactions in traffic has exploded - but often w/o clear underpinning conceptual theory. We review a bunch of theoretical perspectives on interactions, and propose a conceptual framework which sort of unifies them all.
https://psyarxiv.com/8w9z4

Reply on Twitter 1221542578035810307 Retweet on Twitter 12215425780358103073 Like on Twitter 122154257803581030712 Twitter 1221542578035810307

Retweet on Twitter Arkady Zgonnikov Retweeted

Eoin Travers @TraversEoin·

23 Jan

I'm proud in to introduce evidently, a python package for simulating from evidence accumulation models.
https://github.com/EoinTravers/Evidently/
Unlike other packages, it focuses on the time course of these models, and should be particularly useful for looking at their EEG correlates!

Reply on Twitter 1220306317052665856 Retweet on Twitter 122030631705266585637 Like on Twitter 1220306317052665856101 Twitter 1220306317052665856

Retweet on Twitter Arkady Zgonnikov Retweeted

Golan Levin @golan·

17 Dec

My student @LingdongH has released 文言, or wenyan, an esoteric programming language that closely follows the grammar and tone of classical Chinese literature. There's an online IDE, and it compiles to JS or Python. Github: https://github.com/LingDong-/wenyan-lang Site: http://wenyan-lang.lingdong.works/

Reply on Twitter 1206959246988005379 Retweet on Twitter 12069592469880053791132 Like on Twitter 12069592469880053792918 Twitter 1206959246988005379