Artificial Intelligence (former Big Data)
Please take few reflection minutes and try to answer yourself. Just after that read this text further.
How society and science can benefit from AI?
Which are AI downsides?
Which are AI limitations and how they can be overpassed?
Do I really like working on AI or is it just the societal hype?
Take few more reflections moments.
One common approach would be to define AI as the study of intelligent agents. Agents can be anything, from devices to pieces of software. They can learn from data/environment/experience to maximize their chances of succeeding in a goal. Reflection moment: Are these really universal truths? Traditional AI problems involve reasoning, planning, learning, perception, pattern and image recognition, etc. There is a number of subfields in AI. Within this honors track, the main focus will be on deep learning and neural networks, but also other subfields (e.g. machine learning, symbolic reasoning, evolutionary computation) and related fields (e.g. complex systems, neuroscience, optimization) will be addressed. The work will be performed in multidisciplinary teams, following the problem-based and design-based learning approaches. All teams will focus on the technical and scientific aspects of AI, except one which will focus on AI philosophical aspects. All teams will interact with each other.
From September to November, we will try to have a very fast learning curve. This assumes several short lectures on key areas of AI given by experts and by yourself using the flip-classroom concept. The lectures will be far of covering everything, but shall be enough to start creating a basic understanding with many missing elements. Mainly by self-study and with some guidance from coaches you shall be able to fill the dots yourself. The remaining of the year will focus on deeper understanding of AI through practice in real-world competitions (e.g. Kaggle competitions).
In the second year, there will be two main directions based on your personal interests. An applied research direction in which we will participate in data-driven competitions at top level conferences (e.g. NeurIPS Competition Track), and a fundamental research direction. The latter will start in the first semester by entering in the ICLR reproducibility challenge (or similar activities). It will continue in the second semester by developing your own publishable research.
AI is a broad multidisciplinary field. During this two-year track will be impossible to cover all its aspects, but you shall be able to “learn how to learn” yourself about any domain and to understand better what type of career you would like to follow after your bachelor studies (e.g. engineering, management , research, academia). Everyone can find her/his place in this honors track. You need just to be very self-motivated, to have a strong wish to succeed, to like self-study and team work at the same time, and to enjoy answering to questions like “How?” and “Why?”.