Proposals are hereby requested for projects in Human Language Technologies (HLT) for the official South African languages. Funding is available for projects over a period of two to three years. It is foreseen that two to three projects can be funded from this budget. As a general guideline, a project can receive between R1 million and R2,5 million per annum.

DAC

Proposals that address some of the following will be evaluated positively:

  • Projects that are complimentary to the Use of Official Languages Act (Act 12 of 2012).
  • Projects to develop applications with the potential to make a considerable impact, e.g. in a government context. DAC is also interested in applications for mobile devices.
  • Projects that develop resources that contribute to the Basic Language Resource Kits (BLaRKs) for South African languages. 50% of the available funding will be earmarked for language resource development projects. Projects that combine resource and application development will also be viewed favourably.

Factors that will be considered when evaluating the proposals include the following:

  • The scientific and/or development track record of the applicant(s);
  • The practical importance and potential impact of the proposed project in terms of its support of the implementation of the Use of Official Languages Act;
  • The potential contribution of the proposed project to the development of HLTs for all the official languages of South Africa;
  • The extent to which the project commits to releasing at least the basic resources (e.g. language resources and core technologies) developed as part of the project under an open source licence to the Resource Management Agency (www.rma.nwu.ac.za).
  • The alignment of the project to the Intellectual Property Rights from from Publicly Financed Research and Development Act No. 51 of 2008 (“IPR Act”), in terms of the applicant’s intentions to commercialise and/or utilise the project’s outputs to the benefit of the citizens of South Africa.
  • The incorporation of international standards and best practices in the project.
  • The project schedule must provide for evaluation of the project outputs by an independent party to ensure quality and adherence to international best practices. Recommendations for improvement or corrections following such evaluation must be completed before the project can be considered completed.
  • Value for money. Where overheads, indirect costs or hourly rates are considered to be excessive, the Department may opt to offer only partial financial assistance.

Proposals should clearly state the approach that will be taken, and the following should be highlighted:

  • Tasks to be undertaken;
  • Milestones and deliverables;
  • Licensing: details of outputs that will be released to the RMA to be distributed under an open source licence, and outputs that will not be released into the public domain.
  • Costing and budget;
  • Management and governance; and
  • Evaluation criteria and evaluation procedure.

Successful applicants will be required to enter into a formal contract with the Department of Arts and Culture.

The Department of Arts and Culture reserves the right to, within reason, request changes to shortlisted proposals to better align with its strategy, medium-term expenditure framework, etc. Such requests may be directed at, amongst others, specifications, deliverables, budgets and evaluation criteria.

The Department is not bound to accept any of the proposals submitted and reserves the right to annul the call process and reject all proposals at any time prior to contract award.

Closing date for Proposals: Close of business on Monday, 14 December 2015. Contact person (enquiries and submission of proposals):

  • Ms Ulrike Janke at ulrike.janke@dac.gov.za / 082 677 2624

ADDITIONAL INFORMATION

1. WHAT IS HUMAN LANGUAGE TECHNOLOGY?

HLT refers to technologies that process human language. It is the product of a merger between language and computer technology and brings together experts from a wide range of disciplines, including Linguistics, Engineering and Programming. The aim of HLT is to come up with technological solutions to language-related problems.

The Department of Arts and Culture has embraced HLT as part of its strategy to achieve true multilingualism and to elevate the status of the previously marginalised languages as set out in section 6 of our Constitution. As such, the DAC has been granting financial support to HLT projectsw for close to a decade to help grow a local HLT industry that can develop technologies previously reserved for the dominant world languages.

2. WHAT IS THE USE OF OFFICIAL LANGUAGES ACT?

The UOLA requires of every national government department, national enterprise and public entity to adopt a language policy. It is to select at least three official languages in which to communicate with the public, and to set up a language unit to do the required translation work.

The objectives of the UOLA, as stated in clause 2, are:

  1. (a)  to regulate and monitor the use of official languages for government purposes by national government;
  2. (b)  to promote parity of esteem and equitable treatment of all official languages of the Republic;
  3. (c)  to facilitate equitable access to services and information of national government; and
  4. (d)  to promote good language management by national government for efficient public service administration and to meet the needs of the public.

As the custodian of this act, the DAC is interested in the role that HLT can play with regards to the implementation thereof. The UOLA is creating a myriad of challenges, including budgetary issues, the recruitment of skilled translators, coping with growing amounts of work, and the lack of domain-specific terminology in all the official languages. Also, the Act requires departments to put in place a strategy for dealing with members of the public who are blind or deaf, or do not understand one of the chosen languages for communication. These challenges create interesting opportunities for HLT.

3. WHAT IS THE INTELLECTUAL PROPERTY RIGHTS FROM FROM PUBLICLY FINANCED RESEARCH AND DEVELOPMENT ACT?

Section 2(1) of this act reads as follows: “The object of this Act is to make provision that intellectual property emanating from publicly financed research and development is identified, protected, utilised and

commercialised for the benefit of the people of the Republic, whether it be for a social, economic, military or any other benefit.”

“utilised and commercialised” in this context should not be interpreted narrowly. On a case-by-case basis it should be determined whether pure commercial exploitation or unrestricted open source release would serve South Africans best. In some cases it is preferable that intellectual property be protected from exploitation by international players, so as to nurture the local economy and ensure affordability for South African users. In other cases an open source release will maximuse uptake. As a general rule of thumb, the DAC would like to see that a project releases reusable language resources and core technologies into the open domain so that other researchers and developers can use them to do other research and products for the benefit of South African languages, while there is an undertaking to exploit final products in some way that ultimately benefits South Africa best.

How a project proposal responds to this act is a very important aspect in the evaluation and awarding of funding in this programme.

4. WHAT ARE BASIC LANGUAGE RESOURCE KITS?

A BLaRK is the basic set of language resources required to do any useful research and development. This includes text and speech corpora and core technologies such as taggers and text-to-speech systems.

For more information on BLaRKS and the South African position, please consult the following articles:

  • Krauwer, S. 2003. The Basic Language Resource Kit (BLARK) as the First Milestone for the Language Resources Roadmap. Internet: http://www.elsnet.org/dox/krauwer-specom2003.pdf
  • Sharma Grover, A., Van Huyssteen, G.B. & Pretorius, M.W. 2009. The South African Human Language Technologies Audit. Internet: https://dl.dropboxusercontent.com/u/13273429/GepubliseerdeArtikels/Article.RES.SharmaGroverVan HuyssteenPretorius2011a.9.9.9.GBVH.2011-06.28.Final.pdf
  • Sharma Grover, A., Van Huyssteen, G.B. & Pretorius, M.W. 2010. An HLT profile of the official South African languages. Internet: http://aflat.org/files/grover.pdf

Regarding the South African HLT Audit, it should be noted that since 2009, the DAC has funded the development of the following: Named-entity recognisers for 10 languages, phrase chunkers for 10 languages, a language identifier for 11 languages, automatically annotated parliamentary data (English), and machine- translation systems for Setswana and Xitsonga.

5. WHAT IS THE RESOURCE MANAGEMENT AGENCY?

The RMA is a convenient online depository point for various types of electronic data of the official languages for research and development purposes in the field of HLT and linguistics. Most of the resources developed with DAC funding can be found in its catalogue and are available for download and use. We will continue to encourage anyone who has similar resources to licence them to the RMA for distribution, in the interest of the development of all our languages.

The RMA can be visited at the following URL: http://rma.nwu.ac.za