Background

Global Pulse (http://www.unglobalpulse.org/about-new) is an innovation initiative of the United Nations Secretary-General on Big Data. Its vision is a future in which big data is harnessed safely and responsibly as a public good. Its mission is to accelerate discovery, development and scaled adoption of big data innovation for sustainable development and humanitarian action. Global Pulse is building a catalytic network of in-country innovation centers, which will test new tools, data and analysis, share what lessons with the global community, mainstream successful approaches for real-time impact monitoring into policymaking. The first Lab has been established in New York (Pulse Lab New York) followed by Jakarta, Indonesia and Kampala, Uganda.                                                                                                                                               

Pulse Lab Kampala (http://www.unglobalpulse.org/kampala) is an inter-agency initiative of the United Nations under the Office of the Resident Coordinator in Uganda. The first of its kind in Africa, it brings together researchers from Government, the UN, non-governmental organizations and private sector to explore the utility of new sources of digital “Big Data” (such as social media, mobile data, online information) and real-time analysis techniques for social development monitoring and program planning.

A data revolution is underway in Uganda and in Africa. There is more data now than ever. New technologies are leading to an exponential increase in the volume and types of data available, creating unprecedented possibilities for informing and transforming society and protecting the environment. There is more data produced in the last 2 years than in all of human history. There is a tremendous potential in new technology; indeed, this is the forefront of the data revolution already well underway in in the private sector.  The new SDGs agenda will have increased demands – and opportunities- for the use of data. Pulse Lab Kampala supports the UN Country team to tap into the revolution obtaining more and better data to support development work and humanitarian action supporting:

  • Enhance Early Warning: Detection of anomalous trends and patterns of events enables early warning of emerging crises and earlier responses to prevent long-term harm;
  • Real-Time Awareness:  An up-to-date picture of trends, hotspots and dynamic changes in population behavior enhances programme planning and monitoring of implementation and;
  • Rapid Impact Monitoring and Evaluation: Timely feedback on the impact of programmes and policies allows for adaptive course correction and accelerated achievement of results.

Radio is a vibrant medium for public discussion in Uganda. Variety shows and phone-ins hosted by community radio stations are popular ways for Ugandans to voice their needs, concerns and opinions, particularly in rural areas. By using technology to capture audio and transcribe the content into text, radio can become a digital big data source. Pulse Lab Kampala is working closely with Stellenbosh and Makerere Universities to develop the technology to conduct radio mining in Uganda. The project will capture radio audio in Ugandan English, Acholi and Luganda and transcribe the content into text. Pulse Lab Kampala is currently analysing around 40 stations in Kampala and will soon expand to other locations in Uganda. Speech recognition pronunciation dictionaries, and language and acoustic models are under development. Details on the project are available at: http://pulselabkampala.ug/radiomining/

Under the guidance of the Pulse Lab Kampala Data Scientist, an in coordination with the project team at Stellenbosch and Makerere Universities, the consultant will be responsible for the technical development and advisory of the radio mining toolkit.

Duties and Responsibilities

The consultant will be technically responsible for developing and advisory on:

Transcription setup

  • Specify transcription format and provide sample transcription (i.e. text files containing a representation of the words spoken in the audio recordings) for a set of reference recordings from Ugandan radio provided by Pulse Lab Kampala;
  • Word spotter setup;
  • Set up a word spotter using South African language models, applied to Ugandan English radio broadcasts;
  • Ugandan English adaptation;
  • Adaptation of South African language models to Ugandan English, using training data provided by Makerere University of Uganda;
  • Extension to vernacular languages;
  • Develop basic models for word spotting with a limited vocabulary in Luganda and Acholi;
  • System evaluation;
  • Evaluation of the performance of the speech recognition technology developed.

Methodology

  • The expert is expected to work in close consultation with the Data Scientist at Pulse Lab Kampala;
  • Technical advise will be available from collaborators at Stellenbosch University;
  • The expert will work with the team at Pulse Lab Kampala;
  • Copyright and ownership - all materials produced and other intellectual property will be solely owned by Pulse Lab Kampala. Any material provided to the expert will remain the property of Pulse Lab Kampala.

Competencies

  • Plans and produces quality results to meet established goals;
  • Generates innovative, practical solutions to challenging situations;
  • Demonstrates substantive and technical knowledge to meet responsibilities and post requirements with excellence;
  • Demonstrates openness to change and ability to manage complexities;
  • Responds positively to critical feedback and differing points of view.
  • Ability to conceptualize, plan and execute innovative ideas;
  • Displays open, co-operative behavior with other team members;
  • Remains calm and in control, and good humoured even under pressure;
  • Responds positively to critical feedback and differing points of view;
  • Sets priorities, produces quality outputs,
  • Meets deadlines and manages time efficiently.

Required Skills and Experience

Education

  • MSc in Computer Science, Software Engineering or related field; Ph.D will be an asset;
  • Postgraduate studies in Speech and Language Processing.

Experience

  • At least 2 years work and professional experience of speech recognition technology, including knowledge of the tools and concepts needed to develop new speech models;
  • Experience of software engineering practices for prototyping, testing and refining new systems;
  • Advanced knowledge of speech recognition software (HTX).

Core skills

  • Ability to conceptualize, plan and execute innovative ideas.

Language Requirements

Excellent command of English.

Evaluation Criteria

The award of the contract shall be made to the individual consultant whose offer has been evaluated and determined as:

  • Responsive/compliant/acceptable, and;
  • Highest Combined Score (based on the 70% technical offer and 30% price weight distribution) where the minimum passing scores of technical proposal is 70%.

Technical Proposal (Maximum 70 points):

  • Interpretation of the assignment and methodology- 10 points;
  • Relevance of experience (professional/work experience)- 60 points.

Financial Proposal (Maximum 30 points)

  • To be computed as a ratio of the Proposal’s offer to the lowest price among the proposals received by UNDP;
  • Short-listed candidates may be requested to participate in an interview with the selection panel.

The consultant is required to upload an electronic application in one single PDF on the jobs website not later than June 30th, 2015. The application should include the following documents/information:

  • Duly accomplished Letter of Confirmation of Interest and Availability using the template provided by UNDP (see Annex II);
  • Updated Personal CV highlighting past experience in similar assignments and with 3 professional references with contact details (email and telephone);         
  • Technical proposal (1/2 page maximum);
  • Financial proposal (in template provided in Annex II) stating an all-inclusive fixed lump sum fee for this assignment in USD, supported by a breakdown of costs. Such total lump sum price must include professional fee, and costs necessary to conduct the assignment such as communication costs, etc. The consultant will be paid against the completion of specific, measurable deliverables as identified in this TOR.

Please note that applications received after the deadline will not be considered. For any further clarifications; please contact; justine.naiga-bagonza@undp.org; Fredah.zawedde@undp.org and diana.nabbanja@undp.org.

For Annexes I & II - Please check on http://procurement-notices.undp.org/ Reference 23352.