Machine Learning Lab

The Machine Learning Lab headed by Dr Marc Deisenroth was launched on the 22nd March.

The vision of the Machine Learning Lab is to develop autonomous decision-making systems, which close the perception-action-learning loop while learning from small amounts of data.

Therefore, the Machine Learning Lab aims to promote and lead scientific advances in data-efficient machine learning, i.e., the ability to learn in complex domains without requiring large quantities of data. Research areas that fall into this category include probabilistic modelling, incorporation of domain or structural prior knowledge, transfer learning, semi-supervised learning, active learning, Bayesian optimization and reinforcement learning.

To support the Lab the DSI is delighted to announce a £100k Seed Fund for pilot projects in the area of Probabilistic Modelling.

 

Call for Proposals in Probabilistic Modelling

Background
Recent efforts in machine learning have addressed the problem of learning from massive amounts data. We now have highly scalable solutions for problems in object detection and recognition, machine translation, text-to-speech, recommender systems, and information retrieval, all of which attain state-of-the-art performance when trained with large amounts of data. In these domains, the challenge we now face is how to learn efficiently with the same performance in less time and with less data. Other problem domains, such as personalized healthcare, robot reinforcement learning, sentiment analysis, and community detection, are characterized as either small-data problems, or big-data problems that are a collection of small-data problems. The ability to learn in a sample-efficient manner is a necessity in these data-limited domains. Collectively, these problems highlight the increasing need for data-efficient machine learning: the ability to learn in complex domains without requiring large quantities of data. Probabilistic modelling is an important part of data-efficient machine learning and statistical inference when data is scarce. Accounting for uncertainty in modelling, prediction of long-term consequences and decision making is critical in these situations. Application areas include healthcare, robotics, optimization, finance and sentiment analysis.

Project criteria
The scheme will support research projects:
1. Between 1-5 months
2. In the research area of probabilistic modelling
3. With a Principle Investigator who be based at Imperial College, has a permanent contract with the College and is a member of the DSI Fellows scheme.
4. Up to £20k per project funding directly incurred costs only. 
5. Staff and non-staff costs can be included

Further information on funding avaliable and eligability criteria can be found in the call guidance.

How to apply
Applications should be emailed to a.ashley-smith@imperial.ac.uk
Applications should have a Principal Applicant based at Imperial and should contain:

  • A completed Data Science Research Institute Seed Fund Application Form which includes:
    • A written case for support
    • A justification for resources
    • Relevant experience of applicant and co-applicants
  • A one page financial summary: applications should provide information on the requested costs using an InfoEd statement approved by the Department. As InfoEds will not be submitted and will remain as draft, an email confirming departmental approval should also be included in the application, however, proposals do not need to be reviewed by Research Services prior to submission
  • CV of the Principal Investigator and Co-Investigators

Deadlines
Deadline for applications is 12 noon (GMT) on Friday 13th April 2018. Decisions are expected by Friday 11th May 2018.

Contacts
Queries related to research field: m.deisenroth@imperial.ac.uk
Queries regarding to funding and application submission: a.ashley-smith@imperial.ac.uk