Go to JKU Homepage
One reason to expect that realizing general AI is within reach can be attributed to advances in the field of machine learning known as Deep Learning. Deep Learning is large deeply-layered artificial neural networks (ANNs) trained by modern learning algorithms on massive data sets. Deep Learning has revolutionized speech recognition, visual object recognition, object detection, language processing, text analysis, and many other areas, including drug discovery and genomics.
LIT AI Lab Research Goals
We Advance AI.
Artificial Intelligence (AI) is currently impacting and revolutionizing almost all fields of science. AI has infiltrated business and society in an unprecedented manner and despite great success stories, most of AI’s enormous potential must still be realized. Current AI technology focuses on very specific tasks and is far removed from being a general Artificial Intelligence (general AI), a machine that could successfully perform any cognitive task by virtue of its sensory perception, previous experience and “world knowledge”. We are tackling this challenge from different perspectives by advancing machine learning, logical reasoning, and computational perception as well as by exploring and taking advantage of the synergies between these three fields.
Another reason to be optimistic is attributed to advances in the field of Logical Reasoning, meaning automatically deducing conclusions that necessarily follow from known facts and stated premises via logical deduction rules. One key technology in logical reasoning is Boolean Satisfiability (SAT) Solving which, despite its intractability, is successfully used to solve hard problems in a wide range of application domains such as planning and scheduling.
A third reason can be attributed to advances in the field of Machine Perception and the processing, filtering, abstraction, and interpretation of sensory information in order to represent, understand, and ultimately act in the environment. Although machine learning plays a central role here, high-level perception processes continue to be guided by world knowledge and expectations shaped by previous learning. Integrating statistical pattern recognition with long-term learning and logical reasoning processes promises to lead to the next wave of break-throughs.
Leading research groups in the three fields collaborate at the LIT AI Lab in order to contribute to the broad-base evolution of Artificial Intelligence and bring AI to the people. Two further research groups complement these research goals:
In this research group, we focus on translating advancements in artificial intelligence in areas such as medicine, healthcare, molecular biology, chemistry and drug design.
AI technology permeates more and more aspects of our daily life. Considering human factors in the development of AI technology is therefore a timely and highly important research topic, which is unfortunately often neglected. In the Human-centered AI group, we elaborate methods and algorithms to create AI technology that serves humans and is tailored to the individual, instead of adopting a purely system-centric perspective. We focus on the design of comprehensive user models, development of personalization approaches, and studies of human-technology interaction. Furthermore, our research aims at uncovering data biases and algorithmic biases, and at ensuring transparency and fairness of machine learning algorithms, in particular of algorithms powering recommendation systems.
