Commercial Projects (project Detail Page)

Auto Completion
Assigned Programmer-> Rodas Solomon
Client-> Infograf LLC. Hong Kong, China
Company Contact-> Babak Tavangar, CEO, Email: btavangar@gmail.com
Project Schedule-> Started on August 2014.Dead Line, January 2015
 
Project Goal and Description: The Project is to enhance the company’s natural language processing/query technology with AI. It focuses on adding a type-ahead/auto-completion feature for the companies query. Rodas is using open-source software called Cleo for the type ahead feature. After Rodas designed and incorporated AI featured system into Cleo, the software will be connected to the company’s database that is based on graph database system called Neo4j.
 
Final output: The System, after completion, will provide different intelligent suggestions on its auto-completion feature based on the user’s input.
 
Automated Classification of Accounting Items
Assigned Programmer-> Eyob Delele
Client-> Noble & Associates Brisbane, Australia Company
Company Contact-> Andrew Noble, Director, Email: futurewa@gmail.com
Project Schedule-> Started on July 2014. Dead Line, February 2015

 

Project Goal and Description: Given predefined taxonomy for classifying accounting items, the Project task is to create a model from training data via AI machine learning. It focuses on adding a type-ahead/auto-completion feature for the companies query. Eyob is using Moses tool, which is OpenCog’s open-source framework for training patterns to machines.Final output:After Eyob has designed and trained the system, it will be deployed to the company’s database and the system will be able to automatically classify new accounting items into their right category without any manual interaction.

Zoodles Recommendation System
Assigned Programmer-> Eskender Tamerat
Client-> Zoodles Limited, Bellevue, Washington, USA Company
Company Contact-> Eddie Monroe, Senior Programer, Email: eddiemonroe@gmail.com
Project Schedule-> Started on March 2014. Dead Line, February 2015

 

Project Goal and Description: The Project Concept is to build a system that recommends the suitable game for kids from Zoodles’s online game store. Eskeneder uses open source recommender system called Oryx/Cloudera and by applying AI featured codes, it will be incorporated with Zoodles’s recommender system enhanced as content filtering recommender.

Final output: Once it is completed and trained the system will recommend games for kids based on previous records of games played by the same kid and the popular preference of other kids with similar personality.

Longevity Research
Assigned Programmer-> Selamawit Yilma and Meseret Destaw
Client-> Stevia First Corporation Yuba City, California, USA Project
Advisor-> ,Mike Duncan, Bioinformatics expert, Email: mjsduncan@gmail.com
Project Schedule-> Started on September 2014. Dead Line, April 2015

 

The project is to design a system based on Moses, which is OpenCog’s open-source framework for training patterns to machines, that is capable of qualitative replication of De Magalhaes’s Next-generation sequencing (NGS) statistical analysis. It can be applied in any Next-generation sequencing implemented in aging research. Using Moses, the programmers will develop a better data analyser, data set evaluation pipeline and validation which will be the key factors to find the substances that potentially target the identified genes in the NGS phase.

Final output: Accurately trained system that will identify the right substances in the gene identifiers of the NGS will be incorporated to Stevia’s longevity research software.

Integrating API to Neo4j
Assigned Programmer-> Esender Tamerat
Client-> Infograf LLC. Hong Kong, China Company
Company Contact-> Babak Tavangar, CEO, Email: btavangar@gmail.com
Project Schedule-> Started on September 2014. Dead Line, November 2014

 

The Project is to integrate Google’s application programming interface (API) to Neo4j.

Final output: Once Eskender upload the refined system to Infografs’s API, it will be able to recommend You tube channels in the Neo4j database.

Tele Health
Assigned Programmer-> Dagmawi Moges and Eden Niguse
Client->
Company Contact->
Project Schedule-> Started on December 2014. Dead Line, March 2015

 

Project Goal and Description: The Project aims to design a back end web service that collects data sent from pill dispensers (Known as MedElert) to servers, provides an interface that enables Pharmacists and Doctors to set the pill taking time for patients remotely and also informs their caretakers through text and voice message if (when) they fail to take their medicine.The project uses nodeJs as server side programming language, expressJs for MVC (mode view controller) framework and on the user interface side the programers use bootstrap giving the application a responsive feature independent of devices and browsers.  Inconsideration of the future possibility which focuses on integrating the project with OpenCog, the programers build the database with a graph database (neo4j), which is friendly  open source software for the OpenCog framework.

Final output: The final output will be a web service that manages medication of patients using the information feed from the Pill Dispenser.

Attention Offer Network (AON)
Assigned Programmer-> Eskender Tamerat and Dagmawi Moges
Client-> This is for sell
Company Contact-> Josh Cowan: jshcwn@gmail.com
Project Schedule-> Started on March 2015. Dead line July 2015

 

Project Goal and Description:The project goal is to create a system which will facilitate a platform for meaningful, emotionally positive, easy to use exchange of services between two and more persons without monetary payment via the internet. The programer will use open source software like NeoRj on database and bootstrap. The application will be built using Django programing language.

Final Output: At the end, Eskender and his team will deliver an interactive website where users can team up in different projects and work together. The web will be able to match users, their interest, and their skill with the right kind of project.

YaNetu Project

Assigned Programmer-> Abel G/Slase, Dagmawi Moges, Dawit Abrahm and Tensae Birhan

Client-> iCog Labs

Project Schedule-> Started on March 2015. Dead line May 2017

 

The project goal is to develop a Teaching Tablet, which will feature a talking teacher avatar and a talking student avatar as well. The intended platform for these products is either Android, or Ubuntu Touch; it will start out with Android, and it will switch to Ubuntu Touch if Android proves problematic. The pilot-testing version, slated for release in 2016, will comprise novel educational software to run on existing low-cost Android tablets. The wide-release version, slated for 2017, will combine eye-friendly e-reader and video touchscreen capabilities, using newly available screen technology or custom dual-screen hardware.

The key components of the project will include

  • Character programming (human-authored code or other content that determines what each avatar actually does, in response to user input)
  • Avatar visualizer (the “game world” or similar in which the avatar is shown, and with which the user visually interacts)
  • Controller (the software that actually runs the character programming via the avatar visualization software, and via the audio proxy)
  • Audio proxy (the software that handles speech input/output, based on the particulars of the hardware device. The controller will send and receive sound files to/from it.)
  • AI Engine (software that provides avatars with intelligence beyond what is specifically programmed in by humans as character programming)

Final Output: At the end of the project, iCog Labs will own an intelligent teaching tablet {brand name, YaNetu} which is uniquely innovative hardware/software device designed as a learning helper for school age children in sub-Saharan Africa.

Soccer Playing Robots

Assigned Programmer-> Abenezer Nigusse, Israel Abebe, Eskender Besrat, Kidus Wendimagegn, Solomon Molla, Daniel Fikadu, and Dagim Sisay

Client-> For Sale

Company Contact->

Project Schedule-> Started on December 2015. Dead Line, April 2016

Project Goal and Description: The Project is to upgrade the current RoboSapien  models’ hardware and software features in ways that will enable them to play soccer autonomously. The upgraded robots will be tried on the iCog Makers Initiative projects.

The Key components of the Project will include:

  • Localization; each robot will identify its location respective to the playing field.
  • Creating a platform through which team behavior could be developed.
  • Enhancing the existing actuators and manipulators.
  • Enhancing the Vision and communication system (camera placement and serial communication) the  for the Android based hacking.

The programers will use C++ and Python as the programing language. OpenCog’s Moses will be integrated for the machine learning part. Open source software like OPenCV and ROS will also be implemented.

Final output: At the end of the project, iCog will own autonomous robots capable of playing soccer with features like self localization, and collective behavior generating algorithms.

Face Tracker

 

Assigned Programmer →  Tesfa Yohannes
Client → Hanson Robotics
Contact Person →  Vytas Krisciunas (vytas@hansonrobotics.com)
Project Schedule → Started on November 2015. Deadline February 2016
Project Goal and Description: The project aims to create a robust image tracker that is able to track elements at time goes by and gives a better bounding box of elements and also gives indication on failing track results. The programmer would use libraries of tools that perform track functionalities and tweak the libraries to give better indication of the internal state. The application would use cpp for better performance and to leverage ROS tools that are already developed.
Final Output: At task completion; The tools would be able to be reconfigurable to be able to run at multiple cameras in different configuration to be able to tweak the performance of the different trackers that are used in the system. And also tools would include different tools that does image projection to and from 2D to 3D space. And implementation of basic tracking of people eyes when deployed on the robot.
Accessing Skeleton Tracker Information
Assigned Programmer →  Wesagn Dawit
Client → Hanson Robotics
Contact Person →
Project Schedule → Started on July 2015. Deadline March 2016
Project Goal and Description: skeleton tracker is the system that tracks the skeleton of a user standing in front of the camera ms kinect or asus depth camera. the skeleton tracker aims to extract the joint information of the tracked user to integrate the gesture recognition and face recognition into it and later to be embedded into Hanson robotics' Eva head. I use python and c++ programming languages with ros to extract information from the skeleton tracker to show the joint positions inside the tracker.
Final Output: Getting the skeleton information of anyone standing on front of camera in real-time and extracting this future for dealing with human computer interaction.

Comments are closed.