Last month, CDT students travelled to Vancouver to present their work at NeurIPS, one of the largest AI conferences. The schedule was packed, with 6 conference papers and 3 workshop papers presented. The CDT is proud to sponsor and support our PhD students as they present their research at the conference, showcasing their hard work and academic excellence on an international stage!
Day 1, Tuesday: No papers were presented, but the expo, tutorials, and careers fair kept everyone occupied.

Day 2, Wednesday:

William Bankes William and his supervisor presented their work on robust data down sampling. Naive approaches to training on a subset of data can cause problems when classes in the dataset are imbalanced, with rare classes becoming even rarer. A direct result of this is the model’s ability to predict on this subset of the data decreases. To address this they proposed an algorithm called REDUCR, which downsamples data in a manner that preserves the performance of minority classes within the dataset. They show REDUCR works across a range of text and image problems achieving state of the art results. The work is available here.

Jake Cunningham also presented his work on 
Reparameterized Multi-Resolution Convolutions for Long Sequence Modelling

 

 

Day 3, Thursday
Yuchen Zhu 

Yuchen’s joint paper with Jialin Yu and Ricardo Silva from UCL Statistical Sciences, Structured Learning of CompositionalSequential Models was presented as part of the main proceedings of NeurIPS 2024. They proposed an explicit model for expressing how the effect of sequential interventions can be isolated into modules, clarifying previously unclear data conditions that allow for the identification of their combined effect at different units and time steps. The paper is here.  Additionally, together with collaborators from MPI Tuebingen, Yuchen presented a paper Unsupervised Causal Abstraction at the Causal Representation Learning workshop. Due to the lack of interpretability of current large blackbox models, they propose a methodology for causally abstracting a large model to a smaller and more interpretable model. In particular, unlike existing methods, their method does not require supervision signals from the smaller model. The paper can be found  here.

David Chanin

David presented a paper along with co-author Daniel Tan, another UCL PhD Student. They find that Contrastive Activation Addition (CAA) steering has mixed results in terms of robustness and reliability. Steering vectors tend to generalise out of distribution when they work in distribution. However, steerability is highly variable across different inputs: depending on the concept, spurious biases can substantially contribute to how effective steering is for each input, presenting a challenge for the widespread use of steering vectors. While CAA is effective on some tasks, other behaviours turn out to be unsteerable. As a result, it is difficult to ensure they will be effective on a given task of interest, limiting their reliability as a general alignment intervention. The paper is available here.

Day 4, Friday
Reuben Adams

Reuben presented his paper (co-authored with his supervisors) on extending a classic theorem in the PAC-Bayes literature to account for arbitrary outcomes, rather than simply correct or incorrect classifications. Their work provides theoretical insight into the generalisation behaviour of neural networks and the different kinds of errors they can make. Their framework can cover not just Type I and Type II errors, but any kind of error that may occur in multi class classification. You can find the paper here.

 

 

Daniel Augusto
Daniel presented a co-authored paper in collaboration with
Getúlio Vargas Foundation for the main conference track. Their work proposes a new solution to streaming variational Bayesian inference using  GFlowNets as a foundation for their methodology. This was the first work that allows high quality variational inference for discrete parameters without requiring the storage of the whole dataset. They believe this work will be useful for applications in genetics, through the inference of phylogenetic trees, for preference learning, and other big-data contexts. Their paper can be read here.

 

Day 5, Saturday 

Oscar Key Oscar, along with co-authors from Graphcore presented a poster at the workshop on Efficient Natural Language and Speech Processing. Their work considers the top-k operation, which finds the largest k items in a list, and investigates how it can be computed as quickly as possible on the parallel hardware commonly used to run AI applications. Top-k is found in many AI algorithms, so it’s useful to make it fast, e.g. a large language model might use it to select the most important parts of a long prompt. Their full paper is available here.

 

Varsha Ramineni 

Varsha presented her research at the Workshop on Algorithmic Fairness through the Lens of Metrics and Evaluation (AFME 2024). This work addresses the challenge of evaluating classifier fairness when complete datasets, including protected attributes, are inaccessible. They propose a novel approach that leverages separate overlapping datasets, such as internal datasets lacking demographic information  and external sources like census data, to construct synthetic test data with all necessary variables. Experiments demonstrate that our approach produces synthetic data with high fidelity and offers reliable fairness evaluation where real data is limited.  Varsha says that she had an incredible experience attending her first NeurIPS and presenting her work and engaging in meaningful discussions throughout the conference was a deeply rewarding experience, providing invaluable feedback and ideas as the work extends. Do reach out to her if you’d like to learn more!

 

 

 

1st:	Romy Williamson	the convergence of perception
2nd:	Reuben Adams	Nook
3rd:	Pedro José Ferreira Moreira	UCL Summer School
4th:	Kai Biegun	In With The New
5th:	Roberta Chissich	Fores Escape

1st:	Adrian Gheorghiu & Pedro Moreira
Joint 2nd:	Lorenz Wolf. Mirgahney Mohamed & Jake Cunningham
4th:	Sierra Bonilla
5th:	Bernardo Perrone De Menezes Bulcao Ribeiro & Roberta Chissich

On Wednesday 15th November, members of all cohorts within the FAICDT visited our Robotics labs at the newly opened UCL East Campus. Prof Dimitrios Kanoulas very kindly offered up his time and expertise in showing our students around both the Marshgate and Pool Street Labs.

Our students were able to interact with the quadra-ped robots and speak with academics and other researchers about their ground breaking work in the field of AI and robotics.

The visit to UCL East’s robotics lab was impressive. We saw the Boston Dynamics robot navigate stairs and avoid obstacles, showcasing the practical applications of these technologies. The most striking moment was seeing a robot execute a backflip, which highlighted the advanced capabilities in robotics. Dimitrios Kanoulas kindly also gave us a tour around the Marshgate building! The location’s calm park setting was a nice change from the main campus environment. – Sierra Bonilla, Cohort 5

 

A great time was had by all, and we are delighted that our students have been able to connect with other researchers in the AI field here at UCL.

It is undeniable at this point that AI is going to radically shape our future. After decades of effort, the field has finally developed techniques that can be used to create systems robust enough to survive the rough and tumble of the real world. As academics we are often driven by curiosity, yet rather quickly the curiosities we are studying and creating have the potential for tremendous real-world impact.

It is becoming ever more important to keep an eye on the consequences of our research, and to try to anticipate potential risks.

This has been the purpose of our AI discussion series that I have organised for the members of the AI Centre, especially for those on our Foundational AI CDT.

I kicked off the discussion series with a talk outlining the ongoing debate over whether there is an existential risk from AI “going rogue,” as Yoshua Bengio has put it. By this I mean a risk of humanity as a whole losing control over powerful AI systems. While this sounds like science fiction at first blush, it is fair to say that this debate is far from settled in the AI research community. There are very strong feelings on both sides, and if we are to cooperate as a community in mitigating risks from AI, it is urgent that we form a consensus on what these risks are. By presenting the arguments from both sides in a neutral way, I hope I have done a small amount to help those on both ends of the spectrum understand each other. You can watch my talk here: https://www.youtube.com/watch?v=PI9OXHPyN8M

Ivan Vegner, PhD student in NLP at the University of Edinburgh, was kind enough to travel down for our second talk, on properties of agents in general, both biological and artificial. He argued that sufficiently agentic AI systems, if created, would pose serious risks to humanity, because they may pursue sub-goals such as seeking power and influence or increasing their resistance to being switched off—after all, almost any goal is easier to pursue if you have power and cannot be switched off! Stuart Russell pithily puts this as “You can’t fetch the coffee if you’re dead.” Ivan is an incredibly lucid speaker. You can see his talk Human-like in Every way? here https://www.youtube.com/watch?v=LGeOMA25Xvc

For some, a crux in this existential risk question might be whether AI systems will think like us, or in some alien way. Perhaps we can more easily keep AI systems under control if we can create them in our own image? Or could this backfire—could we end up with systems that have the understanding to deceive or manipulate? Professors Chris Watkins and Nello Christianini dug into this question for us by debating the motion “We can expect machines to eventually think in a human-like way.” (Chris for, Nello against). There were many, many questions afterwards, and Chris and Nello very kindly stayed around to continue the conversation. Watch the debate here: https://www.youtube.com/watch?v=zWCUHmIdWhE

Separate from all of this is the question of misuse. Many technologies are dual-use, but their downsides can be successfully limited through regulation. With AI it is different: the scale can be enormous and rapidly increased (often the bottleneck is simply buying/renting more GPUs), there is a culture of immediately open-sourcing software so that anyone can use it, and AI models often require very little expertise to run or adapt to new use-cases. Professor Mirco Musolesi outlined a number of risks he perceives from using AI systems to autonomously make decisions in economics, geo-politics, and warfare. His talk was incredibly thought-provoking: You can see his talk here: https://youtu.be/QH9eYPglgt8

This series has helped foster an ongoing conversation in the AI Centre on the risks of AI and how we can potentially steer around them. Suffice to say it is a minefield.

We should certainly not forget the incredible potential of AI to have a positive impact on society, from automated and personalised medicine, to the acceleration of scientific and technological advancements aimed at mitigating climate change. But there is no shortage of perceived risks, and currently a disconcerting lack of technical and political strategies to deal with them. Many of us at the AI Centre are deeply worried about where we are going. Many of us are optimists. We need to keep talking and increase our common ground.

We’re racing into the future. Let’s hope we get what AI has been promising society for decades. Let’s try and steer ourselves along the way.

 

Reuben Adams is a final year PhD student in the UKRI CDT in Foundational AI.

In the exciting realm of Model-based Reinforcement Learning (MBRL), researchers are constantly pushing the boundaries of what robots can learn to achieve when given access to an internal model of the environment. One key challenge in this field is ensuring that robots can perform tasks safely and reliably, especially in situations where they lack prior data or knowledge about the environment. That’s where the work of Sicelukwanda Zwane comes into play.

Background

In MBRL, robots use small sets of data to learn a dynamics model. This model is like a crystal ball that predicts how the system will respond to a given sequence of different actions. With MBRL, we can train policies from simulated trajectories sampled from the dynamics model instead of first generating them by executing each action on the actual system, a process that can take extremely long periods of time on a physical robot and possibly cause wear and tear.

One of the tools often used in MBRL is the Gaussian process (GP) dynamics model. GPs are fully-Bayesian models that not only model the system but also account for the uncertainty in state observations. Additionally, they are flexible and are able to learn without making strong assumptions about the underlying system dynamics [1].

The Challenge of Learning Safely

When we train robots to perform tasks, it’s not enough to just predict what will happen; we need to do it safely. As with most model classes in MBRL, GPs don’t naturally incorporate safety constraints. This means that they may produce unsafe or unfeasible trajectories. This is particularly true during early stages of learning, when the model hasn’t seen much data, it can produce unsafe and seemingly random trajectories.

For a 7 degree of freedom (DOF) manipulator robot, bad trajectories may contain self-collisions.

 

Distributional Trajectory Sampling

In standard GP dynamics models, the posterior is represented in distributional form – using its parameters, the mean vector and covariance matrix. In this form, it is difficult to reason about

about the safety of entire trajectories. This is because trajectories are generated through iterative random sampling. Furthermore, this kind of trajectory sampling is limited to cases where the intermediate state marginal distributions are Gaussian distributed.

Pathwise Trajectory Sampling

Zwane uses an innovative alternative called “pathwise sampling” [3]. This approach draws samples from GP posteriors using an efficient method called Matheron’s rule. The result is a set of smooth, deterministic trajectories that aren’t confined to Gaussian distributions and are temporally correlated.

Adding Safety

The beauty of pathwise sampling [3] is that it has a particle representation of the GP posterior, where individual trajectories are smooth, differentiable, and deterministic functions. This allows for the isolation of constraint-violating trajectories from safe ones. For safety, rejection sampling is performed on trajectories that violate safety constraints, leaving behind only the safe ones to train the policy. Additionally, soft constraint penalty terms are added to the reward function.

Sim-Real Robot Experiments

To put this approach to the test, Zwane conducted experiments involving a 7-DoF robot arm in a simulated constrained reaching task, where the robot has to avoid colliding with a low ceiling. The method successfully learned a reaching policy that adhered to safety constraints, even when starting from random initial states.

In this constrained manipulation task, the robot is able to reach the goal (shown by the red sphere – bottom row) without colliding with the ceiling (blue – bottom row) using less than 100 seconds of data in simulation.

Summary

Sicelukwanda Zwane’s research makes incremental advances on the safety of simulated trajectories by incorporating safety constraints while keeping the benefits of using fully-Bayesian dynamics models such as GPs. This method promises to take MBRL out of simulated environments and make it more applicable to real-world settings. If you’re interested in this work, we invite you to dive into the full paper, published at the recent IEEE CASE 2023 conference.

References

 

1. M. P. Deisenroth and C. E. Rasmussen. PILCO: A Model-based and Data-efficient Approach to Policy Search. ICML, 2011.
2. S. Kamthe and M. P. Deisenroth. Data-Efficient Reinforcement Learning with Probabilistic Model Predictive Control. AISTATS, 2018.
3. J. T. Wilson, V. Borovitskiy, A. Terenin, P. Mostowsky, and M. P. Deisenroth. Pathwise Conditioning of Gaussian Processes. JMLR, 2021.

 

We are delighted to announce the successful conclusion of our recent workshop on Distance-based Methods in Machine Learning. Held at the historical Bentham House on 27-28th of June, the event brought together approximately 60 delegates, including leading experts and researchers from statistics and machine learning.

The workshop showcased a diverse range of speakers who shared their knowledge and insights on the theory and methodology behind machine learning approaches utilising kernel-based and Wasserstein distances. Topics covered included parameter estimation, generalised Bayes, hypothesis testing, optimal transport, optimization, and more.

The interactive sessions and engaging discussions created a vibrant learning environment, fostering networking opportunities and collaborations among participants. We extend our gratitude to the organising committee, speakers, and attendees for their valuable contributions to this successful event. Stay tuned for future updates on similar initiatives as we continue to explore the exciting possibilities offered by distance-based methods in machine learning.

In September 2022 I had the amazing opportunity to participate in workshops in Rennes and in Sophia Antipolis, and in a doctoral symposium in Alicante!

In poster sessions and talks, I have presented my work on Aggregated Kernel Tests which covers three of my papers. The first one is MMD Aggregated Two-Sample Test where the two-sample problem is considered, in which one has access to samples from two distributions and is interested in detecting whether those come from the same or from different distributions. The second is KSD Aggregated Goodness-of-fit Test in which we consider the goodness-of-fit problem where one is given some samples and is asked whether these come from a given model (with access to its density or score function). In the third one, Efficient Aggregated Kernel Tests using Incomplete U-statistics, we propose computationally efficient tests for the two-sample, goodness-of-fit, and independence problems; this last one consists in detecting dependence between the two components of paired samples. We tackle these three testing problems using kernel-based statistics, in such a setting the performance of these tests is known to heavily depend on the choice of kernels or kernel parameters (i.e. bandwidth parameter). We propose tests which aggregate over a collection of kernels and retain test power, we theoretically prove optimality of our tests under some regularity assumptions, and empirically show that our aggregated tests outperform other state-of-the-art kernel-based tests.

I started the month of September by participating in GOFCP 2022, the 5th Workshop on Goodness-of-Fit, Change-Point and related problems, from 2nd to 4th September in ENSAI in Rennes (France). It was extremely interesting to hear about the latest research in this very specific research field which covers exactly the topics I had been working on since the start of my PhD.

I then went to EURECOM in Sophia Antipolis (France) for MLF 2022, the ELISE Theory Workshop on Machine Learning Fundamentals, from 5th to 7th September. Talks and poster sessions covered the theory of kernel methods, hypothesis testing, partial differential equations, optimisation, Gaussian processes, explainability and AI safety.

Finally, I participated in EDS 2022, the ELLIS Doctoral Symposium 2022, hosted by the ELLIS Alicante at the University of Alicante in Spain from 19th to 23rd September. It was an amazing experience to meet so many other PhD students working on diverse topics in Machine Learning. I especially enjoyed the numerous poster sessions which allowed to engage with other students and discuss their current research!

I am extremely grateful to Valentin Patilea, Motonobu Kanagawa and Aditya Gulati for the respective invitations, and to my CDT (UCL CDT in Foundational AI with funding from UKRI) which allowed me to participate in those workshops/symposium!

Tuesday 1^st November was a busy day at the CDT and UCL Centre for Artificial Intelligence with our joint UKRI CDT poster showcase and AI demo event. Together with the UKRI CDT in AI-Enabled Healthcare we put on an event featuring posters, demos, AI art and robots.

The afternoon began with presentations by the CDT centre directors Prof David Barber and Prof Paul Taylor, as well as our industry sponsor Ulrich Paquet from Deepmind. In attendance were students, academics and industry partners, keen to understand what we have been doing and where our research will take us in the future.

We had approximately 40 posters on display, with a further 19 demonstrations of AI by a variety of groups from Vision to Natural Language Processing. Engagement with the poster presenters was high across the board and a wonderful opportunity for our students to engage with others about the work that they have undertaken the last few years.

We were honoured to have the Provost in attendance to witness just how vibrant and stimulating our centres are as part of a dynamic and successful Computer Science department.

The UCL Centre of Artificial Intelligence have been donated a rare 3D generated AI generated painting of a Amedeo Modigliani, which started as a Masters and then PhD project for Dr. Anthony Bouchard and Dr. George Cann and will be displayed at the AI Centre for all to see.

The day ended with a robot display in the Function Space, showcasing the quadrapod robots that our students are working with both at the AI Centre and the soon to be opened UCL East.

It was wonderful to witness all the different ways in which AI is being applied and developed to help solve some of societies greatest needs and to have the opportunity to share the work of our students with a wider audience.

With thanks to those who attended, our students, director David Barber, AI Centre manager Sarah Bentley and the TSG team for their time, patience and support in helping to make this a hugely successful event.

One of the most challenging parts of starting my PhD journey so far, has been adapting my written work from the highly technical works such as academic papers, to a broader audience such as those at multi-disciplinary research conferences. The formal definition of this skill would be in the D domain of the Researcher Development Framework, D2 “Communication and Dissemination” and D3 “Engagement and Impact”.

My first poster presentation was given at Cumberland Lodge, mainly to fellow students in the FAI CDT, but also to other AI focused CDTs around the country. The second referenced here however was to other PhD students and academics in the UCL interaction centre (UCLIC) who are a multi-disciplinary group focused on all avenues of technology and how we interact with it, combining the fields of computer science and psychology.

One of the presentations at the showcase which I found interesting was by Leon Reicherts, who presented his paper “Do Make me Think!”. This is about how to use conversational user interfaces to ask questions to the user, in order to enable “deeper” thinking and more thorough learning to the user. I like to think of this as an interactive version of rubberducking.

I think applying a similar concept to these talks in beneficial. That is that ideally you can inform the audience with sufficient detail about your project in order to have the opportunity to have questions fed back to you at the end which provoke this “deeper thinking”. The difference in audience in these two presentations was crucial in order to perform the first part of this technique correctly.

These presentations have been very beneficial to me in generating new directions for my research, informing me on how well I communicated my research as well as having my peers send me relevant research papers that they see that they think will be useful to me. I can’t stress enough how much this last point has helped me in keeping up to date in my research and in writing my literature reviews, search engines can’t hold a candle to a good network of researchers!

This work was supported by the Royal Academy of Engineering Chairs in Emerging Technology Scheme (CiET1718/14)