What are intelligent systems?

Intelligent systems are technologically advanced machines that perceive and respond to the world around them. Intelligent systems can take many forms, from automated vacuums such as the Roomba to facial recognition programs to Amazon's personalized shopping suggestions. 

Our department focuses on two main areas within intelligent systems: how machines perceive their environment and how those machines interact with that environment.

One way that such systems can perceive their environment is through vision. The study of how computers can understand and interpret visual information from static images and video sequences emerged in the late 1950s and early 1960s. It has since grown into a powerful technology that is central to the country's industrial, commercial, and government sectors. The key factors that have contributed to this growth are the exponential growth of processor speed and memory capacity as well as algorithmic advances.

The field of intelligent systems also focuses on how these systems interact with human users in changing and dynamic physical and social environments. Early robots possessed little autonomy in making decisions: they assumed a predictable world and perfumed the same action(s) repeatedly under the same conditions. Today, a robot is considered to be an autonomous system that can sense the environment and can act in a physical world in order to achieve some goals.

Applications of intelligent systems

Intelligent systems are poised to fill a growing number of roles in today's society, including:

  • Factory automation
  • Field and service robotics
  • Assistive robotics
  • Military applications
  • Medical care
  • Education
  • Entertainment
  • Visual inspection
  • Character recognition
  • Human identification using various biometric modalities (e.g. face, fingerprint, iris, hand)
  • Visual surveillance
  • Intelligent transportation

Challenges in intelligent systems

Research in intelligent systems faces numerous challenges, many of which relate to representing a dynamic physical world computationally.

  1. Uncertainty: Physical sensors/effectors provide limited, noisy and inaccurate information/action. Therefore, any actions the system takes may be incorrect both due to noise in the sensors and due to the limitations in executing those actions.
  2. Dynamic world: The physical world changes continuously, requiring that decisions be made at fast time scales to accommodate for the changes in the environment.
  3. Time-consuming computation: Searching for the optimal path to a goal requires extensive search through a very large state space, which is computationally expensive. The drawback of spending too much time on computation is that the world may change in the meantime, thus rendering the computed plan obsolete.
  4. Mapping: A lot of information is lost in the transformation from the 3D world to the 2D world. Computer vision must deal with challenges including changes in perspective, lighting and scale; background clutter or motion; and grouping items with intra/inter-class variation.

Studying intelligent systems

Students who want to study intelligent systems will need to be able to understand and integrate knowledge from various subject areas including:

  • Programming
  • Data structures
  • Algorithms
  • Pattern recognition
  • Machine learning
  • Artificial intelligence
  • Physics
  • Numerical methods
  • Psychology

In addition, math skills are very important. You will use trigonometry, linear algebra, and calculus on a regular basis. Statistics and probability are also fundamental skills for all intelligent systems disciplines. Getting proficient with Linux is good to do as well.

Internship & employment opportunities in intelligent systems

Currently there is strong industry demand for people who understand intelligent systems technology and know how to apply it to real-world problems. Graduates in this area can work in academia, national and government labs, and industry companies such as Google, Microsoft, Intel, IBM).