Anil Anthony Bharath

Youngmok Ha, Viktor Schlegel, Yidan Sun et al.

While Local Differential Privacy (LDP) serves as a foundational primitive for distributed data collection, its stringent noise injection requirement often leads to severe degradation in data utility. This degradation stems from the task-agnostic nature of conventional LDP mechanisms, which inject noise uniformly across all dimensions regardless of their relative importance to the downstream objective. To address this issue, we propose a novel approach that mitigates noise in task-relevant subspaces of the data representation. Our method identifies task-critical subspaces via the Jacobian matrix of the public downstream model, selectively attenuates noise along those dimensions, and reshapes the isotropic noise of standard LDP into an anisotropic distribution. This method preserves the uniform per-dimension privacy budget while heterogeneously modulating noise impact across dimensions, thereby substantially enhancing data utility. Furthermore, our approach generalizes to both linear and non-linear models and integrates seamlessly with existing mechanisms. Extensive experiments on CIFAR-10-C (Brightness corruption at the highest severity level 5) demonstrate that integrating our approach improves the utility of PrivUnit2 and PrivUnitG by approximately 20\% at $ε=7.5$. The source code is available at \url{https://github.com/ymha/jacobian-anr-ldp}.

Christoforos Galazis, Anil Anthony Bharath, Marta Varela

Functional analysis of the left atrium (LA) plays an increasingly important role in the prognosis and diagnosis of cardiovascular diseases. Echocardiography-based measurements of LA dimensions and strains are useful biomarkers, but they provide an incomplete picture of atrial deformations. High-resolution dynamic magnetic resonance images (Cine MRI) offer the opportunity to examine LA motion and deformation in 3D, at higher spatial resolution and with full LA coverage. However, there are no dedicated tools to automatically characterise LA motion in 3D. Thus, we propose a tool that automatically segments the LA and extracts the displacement fields across the cardiac cycle. The pipeline is able to accurately track the LA wall across the cardiac cycle with an average Hausdorff distance of $2.51 \pm 1.3~mm$ and Dice score of $0.96 \pm 0.02$.

Jinyue Li, Yuci Liang, Qiankun Li et al.

Computational pathology demands both visual pattern recognition and dynamic integration of structured domain knowledge, including taxonomy, grading criteria, and clinical evidence. In practice, diagnostic reasoning requires linking morphological evidence with formal diagnostic and grading criteria. Although multimodal large language models (MLLMs) demonstrate strong vision language reasoning capabilities, they lack explicit mechanisms for structured knowledge integration and interpretable memory control. As a result, existing models struggle to consistently incorporate pathology-specific diagnostic standards during reasoning. Inspired by the hierarchical memory process of human pathologists, we propose PathMem, a memory-centric multimodal framework for pathology MLLMs. PathMem organizes structured pathology knowledge as a long-term memory (LTM) and introduces a Memory Transformer that models the dynamic transition from LTM to working memory (WM) through multimodal memory activation and context-aware knowledge grounding, enabling context-aware memory refinement for downstream reasoning. PathMem achieves SOTA performance across benchmarks, improving WSI-Bench report generation (12.8% WSI-Precision, 10.1% WSI-Relevance) and open-ended diagnosis by 9.7% and 8.9% over prior WSI-based models.

Yidan Sun, Viktor Schlegel, Srinivasan Nandakumar et al.

Data-driven decision support in high-stakes domains like healthcare and finance faces significant barriers to data sharing due to regulatory, institutional, and privacy concerns. While recent generative AI models, such as large language models, have shown impressive performance in open-domain tasks, their adoption in sensitive environments remains limited by unpredictable behaviors and insufficient privacy-preserving datasets for benchmarking. Existing anonymization methods are often inadequate, especially for unstructured text, as redaction and masking can still allow re-identification. Differential Privacy (DP) offers a principled alternative, enabling the generation of synthetic data with formal privacy assurances. In this work, we address these challenges through three key contributions. First, we introduce a comprehensive evaluation framework with standardized utility and fidelity metrics, encompassing nine curated datasets that capture domain-specific complexities such as technical jargon, long-context dependencies, and specialized document structures. Second, we conduct a large-scale empirical study benchmarking state-of-the-art DP text generation methods and LLMs of varying sizes and different fine-tuning strategies, revealing that high-quality domain-specific synthetic data generation under DP constraints remains an unsolved challenge, with performance degrading as domain complexity increases. Third, we develop a membership inference attack (MIA) methodology tailored for synthetic text, providing first empirical evidence that the use of public datasets - potentially present in pre-training corpora - can invalidate claimed privacy guarantees. Our findings underscore the urgent need for rigorous privacy auditing and highlight persistent gaps between open-domain and specialist evaluations, informing responsible deployment of generative AI in privacy-sensitive, high-stakes settings.

Stathi Fotiadis, Mario Lino, Shunlong Hu et al.

Deep neural networks have become increasingly of interest in dynamical system prediction, but out-of-distribution generalization and long-term stability still remains challenging. In this work, we treat the domain parameters of dynamical systems as factors of variation of the data generating process. By leveraging ideas from supervised disentanglement and causal factorization, we aim to separate the domain parameters from the dynamics in the latent space of generative models. In our experiments we model dynamics both in phase space and in video sequences and conduct rigorous OOD evaluations. Results indicate that disentangled VAEs adapt better to domain parameters spaces that were not present in the training data. At the same time, disentanglement can improve the long-term and out-of-distribution predictions of state-of-the-art models in video sequences.

Tianhong Dai, Hengyan Liu, Kai Arulkumaran et al.

Hindsight experience replay (HER) is a goal relabelling technique typically used with off-policy deep reinforcement learning algorithms to solve goal-oriented tasks; it is well suited to robotic manipulation tasks that deliver only sparse rewards. In HER, both trajectories and transitions are sampled uniformly for training. However, not all of the agent's experiences contribute equally to training, and so naive uniform sampling may lead to inefficient learning. In this paper, we propose diversity-based trajectory and goal selection with HER (DTGSH). Firstly, trajectories are sampled according to the diversity of the goal states as modelled by determinantal point processes (DPPs). Secondly, transitions with diverse goal states are selected from the trajectories by using k-DPPs. We evaluate DTGSH on five challenging robotic manipulation tasks in simulated robot environments, where we show that our method can learn more quickly and reach higher performance than other state-of-the-art approaches on all tasks.

Tianhong Dai, Hengyan Liu, Anil Anthony Bharath

Episodic self-imitation learning, a novel self-imitation algorithm with a trajectory selection module and an adaptive loss function, is proposed to speed up reinforcement learning. Compared to the original self-imitation learning algorithm, which samples good state-action pairs from the experience replay buffer, our agent leverages entire episodes with hindsight to aid self-imitation learning. A selection module is introduced to filter uninformative samples from each episode of the update. The proposed method overcomes the limitations of the standard self-imitation learning algorithm, a transitions-based method which performs poorly in handling continuous control environments with sparse rewards. From the experiments, episodic self-imitation learning is shown to perform better than baseline on-policy algorithms, achieving comparable performance to state-of-the-art off-policy algorithms in several simulated robot control tasks. The trajectory selection module is shown to prevent the agent learning undesirable hindsight experiences. With the capability of solving sparse reward problems in continuous control settings, episodic self-imitation learning has the potential to be applied to real-world problems that have continuous action spaces, such as robot guidance and manipulation.

Tianhong Dai, Kai Arulkumaran, Tamara Gerbert et al.

Deep reinforcement learning has the potential to train robots to perform complex tasks in the real world without requiring accurate models of the robot or its environment. A practical approach is to train agents in simulation, and then transfer them to the real world. One popular method for achieving transferability is to use domain randomisation, which involves randomly perturbing various aspects of a simulated environment in order to make trained agents robust to the reality gap. However, less work has gone into understanding such agents - which are deployed in the real world - beyond task performance. In this work we examine such agents, through qualitative and quantitative comparisons between agents trained with and without visual domain randomisation. We train agents for Fetch and Jaco robots on a visuomotor control task and evaluate how well they generalise using different testing conditions. Finally, we investigate the internals of the trained agents by using a suite of interpretability techniques. Our results show that the primary outcome of domain randomisation is more robust, entangled representations, accompanied with larger weights with greater spatial structure; moreover, the types of changes are heavily influenced by the task setup and presence of additional proprioceptive inputs. Additionally, we demonstrate that our domain randomised agents require higher sample complexity, can overfit and more heavily rely on recurrent processing. Furthermore, even with an improved saliency method introduced in this work, we show that qualitative studies may not always correspond with quantitative measures, necessitating the combination of inspection tools in order to provide sufficient insights into the behaviour of trained agents.

Marta Sarrico, Kai Arulkumaran, Andrea Agostinelli et al.

Deep networks have enabled reinforcement learning to scale to more complex and challenging domains, but these methods typically require large quantities of training data. An alternative is to use sample-efficient episodic control methods: neuro-inspired algorithms which use non-/semi-parametric models that predict values based on storing and retrieving previously experienced transitions. One way to further improve the sample efficiency of these approaches is to use more principled exploration strategies. In this work, we therefore propose maximum entropy mellowmax episodic control (MEMEC), which samples actions according to a Boltzmann policy with a state-dependent temperature. We demonstrate that MEMEC outperforms other uncertainty- and softmax-based exploration methods on classic reinforcement learning environments and Atari games, achieving both more rapid learning and higher final rewards.

Andrea Agostinelli, Kai Arulkumaran, Marta Sarrico et al.

Recently, neuro-inspired episodic control (EC) methods have been developed to overcome the data-inefficiency of standard deep reinforcement learning approaches. Using non-/semi-parametric models to estimate the value function, they learn rapidly, retrieving cached values from similar past states. In realistic scenarios, with limited resources and noisy data, maintaining meaningful representations in memory is essential to speed up the learning and avoid catastrophic forgetting. Unfortunately, EC methods have a large space and time complexity. We investigate different solutions to these problems based on prioritising and ranking stored states, as well as online clustering techniques. We also propose a new dynamic online k-means algorithm that is both computationally-efficient and yields significantly better performance at smaller memory sizes; we validate this approach on classic reinforcement learning environments and Atari games.

Fatmatulzehra Uslu, Anil Anthony Bharath

Demographic studies suggest that changes in the retinal vasculature geometry, especially in vessel width, are associated with the incidence or progression of eye-related or systemic diseases. To date, the main information source for width estimation from fundus images has been the intensity profile between vessel edges. However, there are many factors affecting the intensity profile: pathologies, the central light reflex and local illumination levels, to name a few. In this study, we introduce three information sources for width estimation. These are the probability profiles of vessel interior, centreline and edge locations generated by a deep network. The probability profiles provide direct access to vessel geometry and are used in the likelihood calculation for a Bayesian method, particle filtering. We also introduce a geometric model which can handle non-ideal conditions of the probability profiles. Our experiments conducted on the REVIEW dataset yielded consistent estimates of vessel width, even in cases when one of the vessel edges is difficult to identify. Moreover, our results suggest that the method is better than human observers at locating edges of low contrast vessels.

Kai Arulkumaran, Marc Peter Deisenroth, Miles Brundage et al.

Deep reinforcement learning is poised to revolutionise the field of AI and represents a step towards building autonomous systems with a higher level understanding of the visual world. Currently, deep learning is enabling reinforcement learning to scale to problems that were previously intractable, such as learning to play video games directly from pixels. Deep reinforcement learning algorithms are also applied to robotics, allowing control policies for robots to be learned directly from camera inputs in the real world. In this survey, we begin with an introduction to the general field of reinforcement learning, then progress to the main streams of value-based and policy-based methods. Our survey will cover central algorithms in deep reinforcement learning, including the deep $Q$-network, trust region policy optimisation, and asynchronous advantage actor-critic. In parallel, we highlight the unique advantages of deep neural networks, focusing on visual understanding via reinforcement learning. To conclude, we describe several current areas of research within the field.

Antonia Creswell, Anil Anthony Bharath

Unsupervised learning is of growing interest because it unlocks the potential held in vast amounts of unlabelled data to learn useful representations for inference. Autoencoders, a form of generative model, may be trained by learning to reconstruct unlabelled input data from a latent representation space. More robust representations may be produced by an autoencoder if it learns to recover clean input samples from corrupted ones. Representations may be further improved by introducing regularisation during training to shape the distribution of the encoded data in latent space. We suggest denoising adversarial autoencoders, which combine denoising and regularisation, shaping the distribution of latent space using adversarial training. We introduce a novel analysis that shows how denoising may be incorporated into the training and sampling of adversarial autoencoders. Experiments are performed to assess the contributions that denoising makes to the learning of representations for classification and sample synthesis. Our results suggest that autoencoders trained using a denoising criterion achieve higher classification performance, and can synthesise samples that are more consistent with the input data than those trained without a corruption process.

Antonia Creswell, Anil Anthony Bharath

Generative adversarial networks (GANs) learn to synthesise new samples from a high-dimensional distribution by passing samples drawn from a latent space through a generative network. When the high-dimensional distribution describes images of a particular data set, the network should learn to generate visually similar image samples for latent variables that are close to each other in the latent space. For tasks such as image retrieval and image classification, it may be useful to exploit the arrangement of the latent space by projecting images into it, and using this as a representation for discriminative tasks. GANs often consist of multiple layers of non-linear computations, making them very difficult to invert. This paper introduces techniques for projecting image samples into the latent space using any pre-trained GAN, provided that the computational graph is available. We evaluate these techniques on both MNIST digits and Omniglot handwritten characters. In the case of MNIST digits, we show that projections into the latent space maintain information about the style and the identity of the digit. In the case of Omniglot characters, we show that even characters from alphabets that have not been seen during training may be projected well into the latent space; this suggests that this approach may have applications in one-shot learning.

Antonia Creswell, Kai Arulkumaran, Anil Anthony Bharath

We focus on generative autoencoders, such as variational or adversarial autoencoders, which jointly learn a generative model alongside an inference model. Generative autoencoders are those which are trained to softly enforce a prior on the latent distribution learned by the inference model. We call the distribution to which the inference model maps observed samples, the learned latent distribution, which may not be consistent with the prior. We formulate a Markov chain Monte Carlo (MCMC) sampling process, equivalent to iteratively decoding and encoding, which allows us to sample from the learned latent distribution. Since, the generative model learns to map from the learned latent distribution, rather than the prior, we may use MCMC to improve the quality of samples drawn from the generative model, especially when the learned latent distribution is far from the prior. Using MCMC sampling, we are able to reveal previously unseen differences between generative autoencoders trained either with or without a denoising criterion.

Antonia Creswell, Anil Anthony Bharath

Generative Adversarial Networks (GAN) are able to learn excellent representations for unlabelled data which can be applied to image generation and scene classification. Representations learned by GANs have not yet been applied to retrieval. In this paper, we show that the representations learned by GANs can indeed be used for retrieval. We consider heritage documents that contain unlabelled Merchant Marks, sketch-like symbols that are similar to hieroglyphs. We introduce a novel GAN architecture with design features that make it suitable for sketch retrieval. The performance of this sketch-GAN is compared to a modified version of the original GAN architecture with respect to simple invariance properties. Experiments suggest that sketch-GANs learn representations that are suitable for retrieval and which also have increased stability to rotation, scale and translation compared to the standard GAN architecture.

Kai Arulkumaran, Nat Dilokthanakul, Murray Shanahan et al.

In this paper we combine one method for hierarchical reinforcement learning - the options framework - with deep Q-networks (DQNs) through the use of different "option heads" on the policy network, and a supervisory network for choosing between the different options. We utilise our setup to investigate the effects of architectural constraints in subtasks with positive and negative transfer, across a range of network capacities. We empirically show that our augmented DQN has lower sample complexity when simultaneously learning subtasks with negative transfer, without degrading performance when learning subtasks with positive transfer.