Machine Learning to Generate New Biological Understanding
Call status: Closed
Previous call: 25 January 2017 - 27 April 2017
Funded through responsive mode
Machine learning is a powerful technique that allows the construction of algorithms that can learn from data and make predictions. As such, it has current and future potential to play a crucial role in one of the pressing problems in current biological research, namely how to interpret increasingly large and complex datasets. This may also include extracting information from datasets that may be ambiguous or contain significant noise. The potential applications to biological sciences are broad, with the techniques being increasingly applied to both numerical and image datasets.
Machine Learning is part of the Data Driven Biology strategic priority (see related links).
Effective machine learning requires the development of suitable well described and annotated datasets, at a scale sufficient to train algorithms. This is a non-trivial task. All applications within the highlight should clearly describe the development of these datasets, and how they will be subsequently made available to other researchers. Where appropriate and justifiable, this may be the main focus of the application.
In addition to the above, applications should seek to utilise machine learning techniques to derive new biological knowledge by:
- Quantifying and analysing image datasets or identifying patterns in raw biological data or images
- Adapting and validating existing techniques previously used to address comparable problems, to further biological research
How to apply
This call is closed to applications.
Applicants interested in accessing the highlight should email Michael Ball (see contact below) with an Expression of Interest (EOI) by 24 February 2017, 16:00 describing their proposed research topic and the fit to the call. EOI should be up to two pages of A4, submitted as PDF or as a Word document.
Applicants submitting an EOI will then be contacted by early March 2017 to inform them whether their proposal is appropriate for this highlight. Applicants with applications appropriate to the highlight will submit through the current responsive mode round (17RM2, application deadline 27 April 2017) and their applications will be assessed at the Research Committee meeting on 19-20 September 2017.
We ran a short webinar prior to the close of the EOI phase where we discussed the purpose and scope of the call. The webinar was on 13 February from 10:30.
Although machine learning can be considered as a subset of both artificial intelligence, and computational statistics, this highlight is specifically seeking to exploit advances in machine learning, rather than looking at using alternative, related technologies.
However, application relevant to biological problems in these broader areas are welcome in responsive mode under the Data Intensive Bioscience priority.
Applications that make use of existing well-developed training datasets or exemplar data within our remit are welcome within responsive mode, but will not be considered under this highlight.
We recognise the highly interdisciplinary nature of these projects, and we welcome projects lead by researchers from outside the biological sciences.
We also note the particular relevance to industry, and welcome appropriate industry partnerships which may include ‘stand-alone LINK’ awards or Industrial Partnership awards (see related links).
Given the importance of well described datasets of suitable volume in the development of machine learning techniques, applicants must give appropriate consideration to making these available in line with the BBSRC data sharing policy (see related links). We are happy to discuss any questions that applicants may have regarding sharing data.
We note the significant contribution of staff such as Research Software Engineers (see external links) to interdisciplinary computational projects such as machine learning, and supports recognition of their contributions and encourages applicants to cost them appropriately on applications to this highlight.