The following applications were contributed by members of euCognition in response to a questionnaire. If you haven't completed the questionnaire, please consider doing so.

The applications are listed in the order in which they were submitted.

Playing the childhood game of hide-and-seek.

Applications which includes self-learning, structuring the sensory input, build meta-representations. Applications in the robotics or surveillance domain would be good areas to demonstrate the difference between a system with and without cognition.

Computer game agents able of cunning and deceiving, i.e. agents that can outsmart their opponents to win a game.

Autonomous systems.

Autonomous robots for space exploration, vaccum cleaning, entertainment, lawn mowing, foraging, etc.

Autonomous sub-goals emergency during a goal-directed task.

Multiscale model-based control.

If a system is capable of proposing a new methodology for solving the same kind of problems that was trained to solve in another way, then it shows a distinctive character of cognition. For example, once being trained to recognize a face by colors or lines... the system investigate new techniques of recognition like typical movement of eyes and mouth.

Autonomous robot.

Cognitive prosthesis (for elderly or disabled people).

Robot Guide Dog.

Autonomous Artificial Cell.

In the robotics domain, showing a flexible remapping of senory events on motor responses based on contextual information, memory and prediction.

Controlling situated real-time systems that have multiple conflicting goals and redundant degrees of freedom / mechanisms of achieving single goals. Robots, intelligent virtual agents, intelligent interactive environments, models of primates or other mammals.

Robotic application showing the ability to perform a completely new task and process by perceiving, understanding the tasks, environment and context itself (but not a mobile navigation application)

Accident prevention aids for human drivers, healthcare advisors, games (eg. chess, etc) players

Any type of user interface, any application domain that requires a user interface.

Domains requiring anticipation and (eventually) active construal of subjective meaning (relevance). Social domains demonstrating the development of shared=objectified "knowledge" and the capability of connecting (grounding) communicated knowledge by integrating it into the active information processes.

A robot (not necessarily anthropomorphic) learning to drive a car.

A household assistant for elderly performing transporation tasks.

The ability to "survive" or maintain existence in a dynamic environment

An example of "primitive" cognition (here embodied): a robot that learns a concept such as "if I want to use this knife for cutting, I ought to grasp it at the handle" (as opposed to the blade, which might be the best choice for putting it into the dishwasher). More "complex" (elusive?) cognition (here not embodied): a system that can learn to do basic symbolic and integer math (addition, multiplication and their precedence) discovers by itself the distributive law.

A chess-playing program that plays a good game across a wide range of opponents, without any operator modification of the program.

Sensors, actuators, and a control algorithm for a set of machines in a factory which can maintain certain setpoints, across a wide range of modifications to those machines, without any operator modification of the sensors, actuators, and control algorithm.

A robot arm, sensors, and control algorithm that can pick up any graspable object left in its workspace, across a wide range of workspaces and objects, without any operator modification of the arm, sensors, or algorithm.

A robot that can perform tasks demonstrated to it, across a wide range of tasks, without any operator modification of the robot.

A self-modifying assembly and associated control algorithm, which acts to gather energy by growing solar panels, sensors, and actuators, and which works effectively over a wide range of enviroments without operator modification of the initial assembly or algorithm.

A robot sniffer/rescue dog, a smart vacuum cleaner, an insightful surveillance system even!

Human-robot/computer situated (multimodal) dialogue.

The ability to link learned procedures in a constructive way and to perform (for example ) an unlearned, goal oriented combinatorial task

A virtual avatar who explains the developments within a scene using natural language sentences.

A simulation environment within which synthetic characters interact themselves and with their context, e.g. by means of gestures, voice and behaviour.

A robot system capable of anticipation

A system that can prompt old people with poor short-term memory when they forget what they are doing.

Driver assist systems, autonomous navigation, systems that can model intentions, real-time realistic face/body animations, cognitive rehabiliation systems

Dr. Who's K9

An application that learned a new task from scratch (either by itself or through instruction) - as opposed to the traditional approach of "performance improvement" on a task pre-programmed by the designer.

Intelligent agents

Image interpretation systems that would be able to learn, navigate, reason, revise knowledge, based on both visual information extracted from the image and symbolic knowledge and other heterogeneous information

Cognitive extensions or cognitive prosthetic systems for the disabled (e.g., a summarizing reading machine for the visually-impaired that would let them quickly know if they want to read more).

Intelligent brain-computer interfaces for disabled and paralysed persons.

Feature and macro extraction that convert POMDP to MDP for (larger than) two agent learning scenarios -- the emergence of language.

Finally, here is a commentary by Aaron Sloman on this exercise.

The question (''What applications best demonstrate the distinctive character of cognition'') presupposes an answer that I refused to give to the question about [[Definitions of Cognition]]. However the task of producing an analytical survey of types of environments, types of competences, types of architectures, forms of representation, varieties of learning, etc. that are capable of supporting the different competences in different environments can help us not only to understand biological phenomena but also to produce a general overview of varieties of practical problems and tradeoffs between alternative solutions. For example, there's a difference between design requirements for a type of robot all of whose instances will work in the same sort of environment (e.g. different factories making the same kind of car) and a type of robot whose working environment will vary and change in ways that the designers cannot predict, e.g. domestic robots to be sold all round the world.