Department of CS newsletter

Deadline for outline: 20-Feb-20
Deadline for full proposal 18-Jun-20
https://epsrc.ukri.org/funding/calls/trustworthy-autonomous-systems-research-nodes-call/

Webinar slides now available.
UK Research and Innovation (UKRI) invites outline proposals for the research nodes as part of the UKRI Trustworthy Autonomous Systems programme. Up to £20.5 million (at 80% fEC) funded through the Strategic Priorities Fund (SPF), is available to support seven research nodes (approximately £3 million UKRI contribution each) for 42 months, starting from 1 October 2020 (fixed start date). Each node will focus on one of the following research topics: trust, responsibility, resilience, security, functionality, verifiability and governance & regulation. Please note that the UKRI Trustworthy Autonomous Systems programme is being delivered in two discrete, interconnected parts, the coordination Hub and seven research nodes. Information on the Hub call can be found here.
The nodes will be separate entities connected to the Hub and they will be expected to collaborate broadly with other nodes, the Hub and the wider autonomous systems community.
Please note that this is a cross-UKRI programme, and it is expected that applications will be multidisciplinary, led by and involving any discipline from across the UKRI portfolio.
A two-stage application process will be used in this call. The first stage requires the applicants to submit an outline proposal. Successful outline applicants will be invited to submit a full proposal.

Deadline: 16 March 2020 9am

https://www.staffnet.manchester.ac.uk/bmh/research/funding-and-applications/funding-opportunities/wellcome-trust/public-engagement/

There are 3 grant schemes which are open to academic staff, post-docs, PS staff and postgraduate students (PhD and PGT) from every faculty. We strongly support cross-faculty and interdisciplinary working.
                                                                                                    
·         Emerging Engagement: if you are new to public engagement – you can apply for up to £500
·         Established Engagement: if you already have experience of public engagement - you can apply for up to £3000 (match funding required)
·         Professional Development: if you want to develop your skills in public engagement - you can apply for up to £2000
 
The purpose of this funding is to facilitate imaginative, high quality and inclusive public engagement which reaches new and diverse audiences. 

Deadline for applications: 07-Feb-20
Date of internship: June - September 2020

BBC audiences expect the best content to be available to them in a single place, personalised to their preferences and interests. Datalab was formed to address these issues, by creating a simpler way to discover content. We are doing this by bringing all of our data together into one place, and by using machine learning to enrich it. As we do this, we become able to match our multimedia content with individuals’ interests and context.  
The aim of the internship is for the chosen candidate to explore experimental and “high-risk” research ideas that address problems encountered by the BBC, using machine learning. Areas of interest include in particular :
1. Recommendation Model Development:
o investigate new approaches  to improving the performance of our recommendation models and
o can we use recent state of the art recommendation approaches (e.g. reinforcement learning,  variational auto-encoders)?
2. Recommendation Model Evaluation:
o Investigate new ways of evaluating the performance of our recommendation models.
o How can we measure develop or combine existing low-level metrics to derive measures that are more meaningful to a public service organisation like the BBC?
o How can we meaningfully connect offline evaluation metrics to online behaviour of our users?
3. Ethical AI: Explore new ways to understand the ethical impact of our algorithms.
o Can we measure bias in our datasets that lead to our algorithms treating different audience segments (e.g. gender, age, ethnicity, religion) inappropriately? How can we address this?
o How do we make our algorithms more transparent and explainable to those who use them?
Chosen candidates will be embedded in the “Datascapes” sub-team that is responsible for recommendation services for all text-based BBC products (News, Sport, and World Service). They will be expected, with the support of their  BBC supervisor, to choose a research project and to present their work both orally to the team and in a written report at the end of their placement. The internship has a stipend of £1500/month.
 
Skills required:
Preferred Current Degree
• Bachelors (3rd year)/Masters/PhD (computer science or similar)
Mandatory knowledge/experience:
• Python programming (experience using the Pycharm IDE ideally)
Preferred knowledge/experience:
• Machine learning algorithms (especially those relating to recommendation)
• Code versioning tools (git/github)
• Data Science/ML  python packages (e.g. Pandas, numpy, tensorflow, pytorch etc.)
Deliverables:
• Proof-of-concept that indicates whether or not the idea explored should be pursued further.
• Oral presentation of the project outcome
• Written report
If successful, the work by the chosen candidate could potentially be used by BBC services in the future. 
Location
BBC Design & Engineering, London W12 7TS
BBC Bristol, BS8 2LR

For more information and to apply for the internships, please email dsi.enquiries@manchester.ac.uk

The next group lunch-time session (Peer Support Group for Working Parents) will be on January 28th, Tuesday, 12:30-13:30 in C18, George Begg Building.

This session will be a peer support session in the Faculty, to share unique caring journeys or simply listen to others with empathy. It is a good place to get out of the work with parenting reflections!

Please save it in your diary and feel free to drop in at any time. Tea and coffee will be provided in B Floor Common Kitchen, please do bring along your packed lunch and your mug.