Canal ADDLabs





Research · lines of research

ADDLabs is involved in research and technology in the fields of artificial intelligence and Human-Computer Interaction. See below the lines of research and technology ADDLabs is developing:

Active Documentation

Active documentation technology is based on the ADD model, created by Professor Ana Cristina Bicharra Garcia, and can be compared to the task apprentice. It involves a computer agent that not only has basic knowledge about a given domain, but is also capable of explaining its line of reasoning and learning as it interacts with users. A specialist uses this agent throughout project development; the agent offers suggestions and explanations and also notifies him of possible inconsistencies.

The result of this designer and apprentice-agent collaboration is a specific model, which documents all steps in the entire development process. We use the term Active Documentation because once the model is developed, managers or other specialists can access project development information or explanations of the expert system. These systems have helped improve the work process of the specialists and ensure preservation of knowledge and methodologies that arise from it. In addition, application of the ADD model in different oil related domains saves time and resources during project development.

HCI

Research in the area of Human-Computer Interaction (HCI) is a part of all ADDLabs projects, as these cooperative systems achieve solution based on work in partnership with users.

Thus, studying the dialogue between players (computer and man) with such different thought models fosters mutual understanding. Human-Computer Interaction also has Web related lines of research, especially on the usability of sites and systems designed for the Internet.

For its HCI research and technology, ADDLabs has a laboratory working with Intelligent CAD, Active Documentation, Multi-agent Systems, Cooperative Interfaces, Knowledge Acquisition and Spatial Reasoning.

This environment is dedicated to building interactive products with intuitive and user friendly interfaces. The laboratory’s evaluations are quantitative and qualitative admeasurements of the level of user acceptance of the interface as it is being used. ADDLabs’ Human-Computer Interaction laboratory conducts tests on usability and predictive evaluations.

The laboratory conducts evaluations in all phases of interface development, using videotapes, audiotapes and logs to document activities throughout the process; these include comments and statements, even keystrokes and clicks.

Today this laboratory has an observation room and two testing rooms, with one of the testing rooms designed for group tasks. This environment is conducive to controlled testing, protected from external factors that could distort test results. The environment features acoustic and visual isolation, separating the observation room and two test rooms with a mirror that only allows evaluators to see test participants.

Neural Networks

ADDLabs also researches and develops systems using Neural Networks technology that mirror learning by the human brain, based on the neuron model, for recognizing, processing and storing knowledge to make it accessible for use. The Neural Networks can be adjusted according to the domain and environment, generating satisfactory responses based on complex relationships between the entries provided.

Data Mining

Data Mining is an interactive process for detecting and extracting patterns in large databases, mining for relationships not visible using conventional statistics techniques. Data mining technology helps identify correlations and extract patterns that were not expected. Based on these, analysis reports are generated about the mass of data and rules of association within them.

Visual Computing

The area commonly known as Visual Computing, which includes Computer Graphics , Artificial Vision, Image Processing and Multimedia, applies technology that has been expanding at an accelerated pace in recent decades. With applications for processing geographic data, robotics data, for interactive CAD and for analyzing medical testing images, Visual Computing research involves generating, processing, analyzing and describing visual data within a computer environment.

Fuzzy Logic

Fuzzy Logic is an alternative to traditional notions of pertinence to a set and Aristotelian logic, moving beyond the idea of true/false duality. Some domains have a gradation between true and false, with regions where more than one hypothesis may be acceptable and where traditional logic does not apply. Fuzzy Logic is an alternative for such cases. Through fuzzy logic, quantifiers like "a lot,” "not much," "some" can be mathematically processed.