Continual Learning and Its Applications

Continual learning (often termed as lifelong learning) is the process through which a learning system efficiently learns multiple tasks in succession, which is an essential requirement for Artificial General Intelligence (AGI). The goal is to ensure that the introduction of new tasks does not degrade the performance of the system on previously learned tasks. Moreover, the ideal scenario would be for the system to utilize the knowledge from earlier tasks to improve the learning process for subsequent ones. One of the major hurdles facing machine learning models, especially deep learning systems, in this realm is the "catastrophic forgetting" phenomenon. Although efforts have been made in the deep learning community to address this, a universally successful solution remains elusive.

The objective of this project is to delve into the intricacies of continual learning within the domain of machine learning, devoid of biologically plausible components. Key aspects to be investigated include:

1. Challenges and Knowledge Gaps: Identifying the prevailing challenges in implementing continual learning, understanding the reasons behind catastrophic forgetting, and seeking innovative methodologies to overcome these obstacles.

2. Development and Verification: Crafting efficient continual learning algorithms, followed by rigorous experimentation to validate the proposed methods.

3. Application Domains: Investigating the real-world domains where continual learning holds significant potential. This could span areas like robotics, where machines need to adapt to new tasks without forgetting prior ones, or in areas like healthcare and finance, where data can be non-stationary and the model needs to evolve with the changing data landscape.

4. Performance Metrics: Establishing metrics to measure the efficacy of continual learning models. This would allow for consistent benchmarking and comparison among various approaches.

5. Potential Cross-domain Synergies: Exploring how solutions in continual learning could be adapted or integrated into other machine learning disciplines and challenges.

For an individual considering this project, it is imperative to possess a solid grounding in mathematics and machine learning. Strong programming skills are also a must, given the likely need to develop, test, and refine algorithms. Previous experience in continual or lifelong learning would be an advantage, though not strictly necessary.

This project presents a formidable challenge, breaking new ground in the ever-evolving field of machine learning. It beckons those who are self-driven, eager to embark on cutting-edge research, and keen to find solutions to the long-standing problem of catastrophic forgetting and the broader challenges of continual learning.

If you are interested in this project, please first visit my research student page: https://staff.cs.manchester.ac.uk/~kechen/ for the required materials and information prior to contacting me.

Eligibility

Applicants should have, or expect to achieve, at least a 2.1 honours degree in UG and a master’s (or international equivalent) in a computer science or Math related discipline.

Funding

At Manchester we offer a range of scholarships, studentships and awards at university, faculty and department level, to support both UK and overseas postgraduate researchers applying for competition and self-funded projects.

For more information, visit our funding page or search our funding database for specific scholarships, studentships and awards you may be eligible for.

Before you apply

We strongly recommend that you contact the supervisor for this project before you apply.

How to apply

Apply online through our website: https://uom.link/pgr-apply

When applying, you’ll need to specify the full name of this project, the name of your supervisor, if you already having funding or if you wish to be considered for available funding through the university, details of your previous study, and names and contact details of two referees.

Your application will not be processed without all of the required documents submitted at the time of application, and we cannot accept responsibility for late or missed deadlines. Incomplete applications will not be considered.

Equality, diversity and inclusion is fundamental to the success of The University of Manchester, and is at the heart of all of our activities. We know that diversity strengthens our research community, leading to enhanced research creativity, productivity and quality, and societal and economic impact.

We actively encourage applicants from diverse career paths and backgrounds and from all sections of the community, regardless of age, disability, ethnicity, gender, gender expression, sexual orientation and transgender status. 

After you have applied you will be asked to upload the following supporting documents:

• Final Transcript and certificates of all awarded university level qualifications

• Interim Transcript of any university level qualifications in progress

• CV

• Contact details for two referees (please make sure that the contact email you provide is an official university/work email address as we may need to verify the reference)

• English Language certificate (if applicable)

If you have any questions about making an application, please contact our admissions team by emailing FSE.doctoralacademy.admissions@manchester.ac.uk.