x Anita Verő

Anita Lilla Verő

Entrepreneur, AI and Content Consultant, PhD in AI, Singer, Writer
anitaveroe _at_ gmail.com


likedin github medium

About me

I'm an Entrepreneur, AI and Content Consultant at ConsonantAI.

I’m also working on a project to develop solutions (including both AI and non-technical ones) to maintain social bonds and mental well-being, called The Third Thing (kick-off article).

Previously, I was a Senior Machine Learning Research Engineer, content writer and manager at Unitary. I worked on building multimodal video undersanding models for online safety and more. I often write about responsible AI and mental health related topics.

Before this, I did a PhD in Multimodal NLP in the Natural Language and Information Processing Group, at the University of Cambridge Computer Laboratory, supervised by Prof Ann Copestake, previously by Dr. Stephen Clark. I am a member of King's College.

I did my undergraduate and masters degrees in computer science at the Eötvös Loránt University. During my masters studies and for one more year I have been working at the Neural Information Processing Group. During this period I spent two months as a guest researcher at the German Research Center for Artificial Intelligence in Saarbrücken, Germany.

When I'm not doing research I am working on music and writing projects, hiking, cycling or doing yoga.

See my CV for a description of my educational background and work experience.


Interests

Reimagining Human Connections in a Globalized, Virtual World -- With families, friends, and colleagues often scattered around the world due to work commitments, maintaining deep, meaningful relationships becomes a challenge. This can lead to widespread feelings of isolation. How could we reimagine our ways of maintaining meaningful relationships using technological and other solutions which help us bring together our offline and online lives?

Responsible AI -- As in the past few years AI technology has become ubiquitous in a large part of the world (and continues conquering more) I have become painfully aware of the complicated and often elusive side-effects of it on society. Beside crucial questions, such as data privacy, fake news, deep fakes and harmful content on the internet, I am preoccupied with the societal consequences of human - AI interaction. We interact on a daily basis with recommendation systems, generative AI (chatbots, image and sound generation). More elusively, AI is already used in sectors such as health care, jurisdiction or insurance which have major systemic impact. Some questions which occupy my mind lately are:

See my blog where I write about these issues as well.

analysis scheme

PhD Thesis -- My PhD research lies in the intersection of machine learning, natural language processing, computer vision and multi-modal semantic grounding. It's focus is on a deeper analysis of data sources and models, unearthing how different meaning representations "understand" concepts. The aim is to make a step towards interpretable models and to move beyond standard evaluation metrics such as accuracy.


Publications


Software

Wikipedia clusters

EmbEval

A Python toolkit for transparent and interpretable evaluation of Embeddings of different sources and modalities.

Twitter income bias

Twitter Global Class Bias

My submission to the first Twitter Algorithmic Bias Challenge. I tested Twitter's image thumbnail cropping the algorithm for images of cheap versus expensive objects, using the income labels as proxy in the dataset of Dollar Street. I found that the algorithm is biased towards cheap rooms and spaces and towards expensive objects. See my report on the detailed findings.

Concept Game

Concept Game

is a two player collaborative game, where players guess the concept for a list of words. It is similar to the famous gamified data collection game ESP Game, but with word clusters instead of images.

The method used to generate the puzzles will be part of my PhD Thesis. Coming soon...

The game has been developed with the help of my friend, Krisztián Gergely.

The code is available on Github

Try Concept Game

Flickr images

MMFeat Flickr API

I developed a Python Flickr API for image search to Douwe Kiela's mmfeat github project. It is a Python toolkit aiming to make it easier for researchers to use multi-modal features. Both image and sound (i.e., visual and auditory representations) are supported.


Teaching / Supervising

Lent 2017 - I was demonstrating and ticking for Machine Learning and Real-world Data, A Part IA CST course at Computer Laboratory, University of Cambridge.

2016/17 - I co-supervised Christopher Davis, a Master student, on his thesis project about mutli-modal semantics. He got distinction and later a PhD position in our research group.


Communal

2017/18 - I was a Mentoring Officer and Representative in the Cambridge Computer Laboratory's Graduate Forum of Women@CL. Women@CL is an organisation to support women in computer science. It involves a mentoring system and provides representation of role models by organising talks and conferences.

2017 - I organised the Cambridge NLIP Seminars, which is a weekly seminar series in NLP and Machine Learning related topics.


Awards


guitar and mic

Music

I have been singing since my childhood, starting as a member of the children choir of the Hungarian Opera House.

Before I moved to the UK I was the lead singer (and occasional guitarist) of an alternative rock band with whom we made some recordings.

You can listen to covers and my original songs on SoundCloud or on my YouTube channel.

poetry

Poetry

I recently attended this poetry writing workshop, which I really enjoyed and learned a lot from.

The poems I wrote for the workshop prompts are available in a mini anthology titled
"The Cassette of Confinement".
It can be downloaded in ebup and PDF formats.

Attribution 4.0 International (CC BY 4.0) Attribution 4.0 International (CC BY 4.0)