EPSRC Network+ proposal: Human-Like Computing
Lead Research Organisation:
Imperial College London
Department Name: Computing
Abstract
Human-Like Computing (HLC) research aims to endow machines with human-like perceptual, reasoning
and learning abilities which support collaboration and communication with human beings. Such abilities
should support computers in interpreting the aims and intentions of humans based on learning and accu-
mulated background knowledge. The Network fits within the EPSRC national priority areas of a) New and Emerging areas, b) People at the Heart of ICT, c) Cross-Disciplinarity and Co-Creation and d) Safe and Secure ICT.
and learning abilities which support collaboration and communication with human beings. Such abilities
should support computers in interpreting the aims and intentions of humans based on learning and accu-
mulated background knowledge. The Network fits within the EPSRC national priority areas of a) New and Emerging areas, b) People at the Heart of ICT, c) Cross-Disciplinarity and Co-Creation and d) Safe and Secure ICT.
Planned Impact
The main stakeholders in our project are: other researchers within both Artificial Intelligence and Cognitive
Science. These stake-holders are drawn from academia, industry and governmental and non-governmental
organisations. We predict that the main impacts will be on the development of a cohesive community of
scientific researchers in the area of Human-Like Computing (HLC). This community will be built around the
development of new theory, implementations and applications of HLC. Such systems will, in the longer term
impact the role of computation within tasks involving close and open-ended interactions between humans
and machines. In particular, a more symmetric interaction between human and machines has the potential
for a closer, more symbiotic relationship between humans and machines in knowledge intensive tasks.
We will communicate directly with our fellow researchers and developers in both academia and industry.
Impact on end users will be indirect: via our contacts in industry, especially with the BBC, Microsoft, Intel
and Syngenta. These companies already have good contact with end users, understanding of their needs
and experience in delivering solutions that meet these needs. It would be an ineffective and inefficient use
of our resources to try to duplicate these contacts, understanding and experience.
Science. These stake-holders are drawn from academia, industry and governmental and non-governmental
organisations. We predict that the main impacts will be on the development of a cohesive community of
scientific researchers in the area of Human-Like Computing (HLC). This community will be built around the
development of new theory, implementations and applications of HLC. Such systems will, in the longer term
impact the role of computation within tasks involving close and open-ended interactions between humans
and machines. In particular, a more symmetric interaction between human and machines has the potential
for a closer, more symbiotic relationship between humans and machines in knowledge intensive tasks.
We will communicate directly with our fellow researchers and developers in both academia and industry.
Impact on end users will be indirect: via our contacts in industry, especially with the BBC, Microsoft, Intel
and Syngenta. These companies already have good contact with end users, understanding of their needs
and experience in delivering solutions that meet these needs. It would be an ineffective and inefficient use
of our resources to try to duplicate these contacts, understanding and experience.
Publications
Ai L
(2021)
Beneficial and harmful explanatory machine learning
in Machine Learning
Cropper A
(2019)
Learning higher-order logic programs
in Machine Learning
Cropper A
(2018)
Learning efficient logic programs
in Machine Learning
Cropper A
(2020)
Turning 30: New Ideas in Inductive Logic Programming
Cropper A
(2020)
Learning Higher-Order Programs through Predicate Invention
in Proceedings of the AAAI Conference on Artificial Intelligence
Hocquette C
(2020)
Complete Bottom-Up Predicate Invention in Meta-Interpretive Learning
Muggleton FREng SH
(2023)
Hypothesizing an algorithm from one example: the role of specificity.
in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
Muggleton S
(2018)
Meta-Interpretive Learning from noisy images
in Machine Learning
| Description | Human-Like Computing (HLC) research aims to endow machines with human-like perceptual, reasoning and learning abilities which support collaboration and communication with human beings. Such abilities should support computers in interpreting the aims and intentions of humans based on learning and accumulated background knowledge. The development of computer systems which exhibit truly human-like computing and co-operative properties will require sustained inter-disciplinary collaboration between disparate and largely disconnected research communities within AI and Psychology. The Network is forging a new UK-based scientific community involving collaboration between leading groups in these disciplines. To date we have published two calls for grants in the area. These were issued in April 2018 and February 2019. The first call led to several successful kick-start grants and a travel grant. A workshop on Third-Wave Artificial intelligence was held at Imperial College in April 2019 followed by the Machine Intelligence 21 Human-Like Computing conference at Cumberland Lodge, in July 2019. A book was published in 2021 by Oxford University Press based on papers presented at the Machine Intelligence 20 and Machine 21 workshops. A Hooke meeting in 2022 on Cognitive Artificial Intelligence has been agreed by the Royal Society, with proceedings published in the Proceedings of the Royal Society B. |
| Exploitation Route | In 2021 we published the book with Oxford University Press. A Royal Society Hooke meeting on Cognitive Artificial Intelligence will be held in area in September 2022. A roadmap document for the area has been developed by Nigel Birch with input from the HLC MI20 community. We are planning to update the roadmap following HLC MI21 in July 2019. The outcomes of each kick-start grant will be reported in the final reports submitted by the associated investigators. A bibliography of associated scientific papers is being built incrementally, and will be made accessible from the website above. Periodically we circulate a newsletter for the network to all members. There have been six issues to date. The Human-Like-Computing workshop has been invited to take part as one of four components in the International Joint Conference on Learning and Reasoning in September 2022. This meeting is associated with a special issue of the Machine Learning Journal, and will involve presentations by network members as well as prestigious plenary invited speakers from both leading tech industries as well as international US universities including MIT and Stanford. |
| Sectors | Agriculture, Food and Drink,Chemicals,Construction,Creative Economy,Digital/Communication/Information Technologies (including Software),Environment,Financial Services, and Management Consultancy,Pharmaceuticals and Medical Biotechnology |
| URL | http://hlc.doc.ic.ac.uk |
| Description | The work of the EPSRC HLC network has led to increasing take-up by the tech industry in work relating to explainable Artificial intelligence. For instance, IBM's Watson Centre in New York held an international conference in January 2022 to which Prof Stephen Muggleton gave a keynote presentation in ongoing work supported by the HLC network on Neuro-Symbolic Learning using Inductive Logic Programming (ILP). Muggleton was introduced at the meeting as the founder of ILP, which is an area which now has dedicated research groups at IBM Watson in New York (led by Dr Alexander Gray), as well as at Google/DeepMind in London (led by Dr Richard Evans) and at Microsoft Redmond in Seattle (led by Dr Summit Gulwani). |
| First Year Of Impact | 2022 |
| Sector | Creative Economy,Digital/Communication/Information Technologies (including Software),Financial Services, and Management Consultancy |
| Impact Types | Societal,Economic |
| Description | Collaboration with A*Star Singapore. |
| Organisation | Agency for Science, Technology and Research (A*STAR) |
| Country | Singapore |
| Sector | Public |
| PI Contribution | Prof Muggleton visited A*Star in October 2018 where he gave an Keynote talk at the main A*Star conference, which was attended by around 700 scientists. This was followed by a week of intensive discussions with research leaders from groups involved in Human-Like collaboration and with the board of A*Star. The discussions also included a panel discussion on Human-Computer collaboration. A group from A*Star is conducting a return visit in March 2019 to continue the discussions. |
| Collaborator Contribution | A*Star leaders each gave presentations of the areas of ongoing research. Potential collaborations and methods were discussed. These could support visits to Singapore by members of staff, post-docs and students. |
| Impact | Kenneth Kwok from A*Star has agreed to give an invited talk at the MI21-HLC (July 2019) meeting hosted by the EPSRC Human-Like Computing Network. The expected outcome is a continuation and expansion of the collaboration. |
| Start Year | 2017 |
| Description | Research partnership with Prof Tenenbaum at Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology. |
| Organisation | Massachusetts Institute of Technology |
| Country | United States |
| Sector | Academic/University |
| PI Contribution | Research visits have occurred in both directions between Imperial College and MIT. In the latest visit Stephen Muggleton and Andrew Cropper were funded by the EPSRC Human-Like Computing network to conduct discussions and plan experiments on identifying relevance of background knowledge for human and machine learners. |
| Collaborator Contribution | Prof Tenenbaum hosted the meetings, which involved the Imperial group giving several presentations to Tenenbaum's group. Around 30 group members attended and contributed to the discussion. Prof Tenenbaum has agreed to give an invited presentation on his latest research at the MI21-HLC meeting being hosted at Cumberland Lodge. Funding for the meeting is from the EPSRC Human-Like Computing network. |
| Impact | D. Lin, E. Dechter, K. Ellis, J.B. Tenenbaum, and S.H. Muggleton. Bias reformulation for one-shot function induction. In Proceedings of the 23rd European Conference on Artificial Intelligence (ECAI 2014), pages 525-530, Amsterdam, 2014. IOS Press. |
| Start Year | 2013 |