Search

This paper investigates the role of communication in improving coordination within robot swarms, focusing on a paradigm where learning and execution occur simultaneously in a decentralized manner. We highlight the role communication can play in addressing the credit assignment problem (individual contribution to the overall performance), and how it can be influenced by it. We propose a taxonomy of existing and future works on communication, focusing on information selection and physical abstraction as principal axes for classification: from low-level lossless compression with raw signal extraction and processing to high-level lossy compression with structured communication models. The paper reviews current research from evolutionary robotics, multi-agent (deep) reinforcement learning, language models, and biophysics models to outline the challenges and opportunities of communication in a collective of robots that continuously learn from one another through local message exchanges, illustrating a form of social learning., Comment: 17 pages, 3 Figures

Modelling partial differential equations (PDEs) is of crucial importance in science and engineering, and it includes tasks ranging from forecasting to inverse problems, such as data assimilation. However, most previous numerical and machine learning approaches that target forecasting cannot be applied out-of-the-box for data assimilation. Recently, diffusion models have emerged as a powerful tool for conditional generation, being able to flexibly incorporate observations without retraining. In this work, we perform a comparative study of score-based diffusion models for forecasting and assimilation of sparse observations. In particular, we focus on diffusion models that are either trained in a conditional manner, or conditioned after unconditional training. We address the shortcomings of existing models by proposing 1) an autoregressive sampling approach that significantly improves performance in forecasting, 2) a new training strategy for conditional score-based models that achieves stable performance over a range of history lengths, and 3) a hybrid model which employs flexible pre-training conditioning on initial conditions and flexible post-training conditioning to handle data assimilation. We empirically show that these modifications are crucial for successfully tackling the combination of forecasting and data assimilation, a task commonly encountered in real-world scenarios., Comment: Accepted at NeurIPS 2024

Training generative models to sample from unnormalized density functions is an important and challenging task in machine learning. Traditional training methods often rely on the reverse Kullback-Leibler (KL) divergence due to its tractability. However, the mode-seeking behavior of reverse KL hinders effective approximation of multi-modal target distributions. To address this, we propose to minimize the reverse KL along diffusion trajectories of both model and target densities. We refer to this objective as the reverse diffusive KL divergence, which allows the model to capture multiple modes. Leveraging this objective, we train neural samplers that can efficiently generate samples from the target distribution in one step. We demonstrate that our method enhances sampling performance across various Boltzmann distributions, including both synthetic multi-modal densities and n-body particle systems., Comment: 23 pages, 6 figures, 3 tables, 1 algorithm

Batched Bayesian optimization (BO) can accelerate molecular design by efficiently identifying top-performing compounds from a large chemical library. Existing acquisition strategies for batch design in BO aim to balance exploration and exploitation. This often involves optimizing non-additive batch acquisition functions, necessitating approximation via myopic construction and/or diversity heuristics. In this work, we propose an acquisition strategy for discrete optimization that is motivated by pure exploitation, qPO (multipoint Probability of Optimality). qPO maximizes the probability that the batch includes the true optimum, which is expressible as the sum over individual acquisition scores and thereby circumvents the combinatorial challenge of optimizing a batch acquisition function. We differentiate the proposed strategy from parallel Thompson sampling and discuss how it implicitly captures diversity. Finally, we apply our method to the model-guided exploration of large chemical libraries and provide empirical evidence that it performs better than or on par with state-of-the-art methods in batched Bayesian optimization.

Current methods for compressing neural network weights, such as decomposition, pruning, quantization, and channel simulation, often overlook the inherent symmetries within these networks and thus waste bits on encoding redundant information. In this paper, we propose a format based on bits-back coding for storing rotationally symmetric Transformer weights more efficiently than the usual array layout at the same floating-point precision. We evaluate our method on Large Language Models (LLMs) pruned by SliceGPT (Ashkboos et al., 2024) and achieve a 3-5% reduction in total bit usage for free across different model sizes and architectures without impacting model performance within a certain numerical precision., Comment: 14 pages, 3 figures

Multi-fidelity Bayesian Optimization (MFBO) is a promising framework to speed up materials and molecular discovery as sources of information of different accuracies are at hand at increasing cost. Despite its potential use in chemical tasks, there is a lack of systematic evaluation of the many parameters playing a role in MFBO. In this work, we provide guidelines and recommendations to decide when to use MFBO in experimental settings. We investigate MFBO methods applied to molecules and materials problems. First, we test two different families of acquisition functions in two synthetic problems and study the effect of the informativeness and cost of the approximate function. We use our implementation and guidelines to benchmark three real discovery problems and compare them against their single-fidelity counterparts. Our results may help guide future efforts to implement MFBO as a routine tool in the chemical sciences.

A coherent concatenation of multiple Townes solitons may lead to a stable infrared and visible broadband filament in ceramic YAG polycrystal. This self-trapped soliton train is leveraged to implement self-referenced multiplex coherent anti-Stokes Raman scattering imaging. Simulations and experiments illustrating the filamentation process and the concatenation of focusing-defocusing cycles in ceramic and crystal YAG are presented. In addition, our simulations and experiments further examine the dependence of the filamentation onset location and supercontinuum generation upon peak input power. Understanding this dependence is key for implementation of viable CARS imaging techniques, due to the comparatively exceptional ability of YAG to generate supercontinuum which can enable higher-sensitivity imaging without delay lines.

Developing an efficient sampler capable of generating independent and identically distributed (IID) samples from a Boltzmann distribution is a crucial challenge in scientific research, e.g. molecular dynamics. In this work, we intend to learn neural samplers given energy functions instead of data sampled from the Boltzmann distribution. By learning the energies of the noised data, we propose a diffusion-based sampler, Noised Energy Matching, which theoretically has lower variance and more complexity compared to related works. Furthermore, a novel bootstrapping technique is applied to NEM to balance between bias and variance. We evaluate NEM and BNEM on a 2-dimensional 40 Gaussian Mixture Model (GMM) and a 4-particle double-well potential (DW-4). The experimental results demonstrate that BNEM can achieve state-of-the-art performance while being more robust., Comment: 20 pages, 7 figures, 2 tables

Diffusion models have shown promising potential for advancing Boltzmann Generators. However, two critical challenges persist: (1) inherent errors in samples due to model imperfections, and (2) the requirement of hundreds of functional evaluations (NFEs) to achieve high-quality samples. While existing solutions like importance sampling and distillation address these issues separately, they are often incompatible, as most distillation models lack the necessary density information for importance sampling. This paper introduces a novel sampling method that effectively combines Consistency Models (CMs) with importance sampling. We evaluate our approach on both synthetic energy functions and equivariant n-body particle systems. Our method produces unbiased samples using only 6-25 NFEs while achieving a comparable Effective Sample Size (ESS) to Denoising Diffusion Probabilistic Models (DDPMs) that require approximately 100 NFEs., Comment: 10 pages, 4 figures

We address the problem of learning uncertainty-aware representations for graph-structured data. While Graph Neural Ordinary Differential Equations (GNODE) are effective in learning node representations, they fail to quantify uncertainty. To address this, we introduce Latent Graph Neural Stochastic Differential Equations (LGNSDE), which enhance GNODE by embedding randomness through Brownian motion to quantify uncertainty. We provide theoretical guarantees for LGNSDE and empirically show better performance in uncertainty quantification., Comment: 9 pages including appendix

Among the recent extensions to standard General Relativity, $f(R,\mathcal{L}_m)$ gravity has risen an interest given the possibility of coupling between geometry and matter. We examine the simplest model with non-minimal coupling in the context of cosmology. We pay special attention to the question of how far this model could reproduce the observational fact of our universe., Comment: 28 pages, 11 figures

Progress in neuroscience has provided unprecedented opportunities to advance our understanding of brain alterations and their correspondence to phenotypic profiles. With data collected from various imaging techniques, studies have integrated different types of information ranging from brain structure, function, or metabolism. More recently, an emerging way to categorize imaging traits is through a metric hierarchy, including localized node-level measurements and interactive network-level metrics. However, limited research has been conducted to integrate these different hierarchies and achieve a better understanding of the neurobiological mechanisms and communications. In this work, we address this literature gap by proposing a Bayesian regression model under both vector-variate and matrix-variate predictors. To characterize the interplay between different predicting components, we propose a set of biologically plausible prior models centered on an innovative joint thresholded prior. This captures the coupling and grouping effect of signal patterns, as well as their spatial contiguity across brain anatomy. By developing a posterior inference, we can identify and quantify the uncertainty of signaling node- and network-level neuromarkers, as well as their predictive mechanism for phenotypic outcomes. Through extensive simulations, we demonstrate that our proposed method outperforms the alternative approaches substantially in both out-of-sample prediction and feature selection. By implementing the model to study children's general mental abilities, we establish a powerful predictive mechanism based on the identified task contrast traits and resting-state sub-networks., Comment: 57 pages, 6 figures

Wave propagation is a common occurrence in all of physics. A linear approximation provides a simpler way to describe various fields related to observable phenomena in laboratory physics as well as astronomy and cosmology, allowing us to probe gravitation through its effect on the trajectories of particles associated with those fields. This paper proposes a unified framework to describe the wave propagation of a set of interacting tensor fields that obey coupled homogeneous linear second-order partial differential equations for arbitrary curved spacetimes, both Lorentzian and metric-affine. We use JWKB Ans\"atze for all fields, written in terms of a perturbation parameter proportional to a representative wavelength among them, deriving a set of hierarchical algebraic and differential equations that link the fields' phases and different order amplitudes. This allows us to reobtain the well-known laws of geometrical optics and beyond geometrical optics in a generalized form, showing that these laws are independent of the rank of the fields involved. This is true as long as what we refer to as the kinetic tensor of a given field satisfies a set of diagonality conditions, which further imply a handful of simplifications on the transport equations obtained in the subleading orders of the JWKB Ans\"atze. We explore these results in several notable examples in Lorentzian and metric-affine spacetimes, illustrating the reach of our derivations in general relativity, reduced Horndeski theories, spacetimes with completely antisymmetric torsion and Weyl spacetimes. The formalism presented herein lays the groundwork for the study of rays associated with different types of waves in curved spacetimes and provides the tools to compute modifications to their brightness evolution laws, consequential distance duality relations, and beyond geometrical optics phenomena., Comment: 24 pages, 2 figures

Bayesian optimization (BO) is a principled approach to molecular design tasks. In this paper we explain three pitfalls of BO which can cause poor empirical performance: an incorrect prior width, over-smoothing, and inadequate acquisition function maximization. We show that with these issues addressed, even a basic BO setup is able to achieve the highest overall performance on the PMO benchmark for molecule design (Gao et al 2022). These results suggest that BO may benefit from more attention in the machine learning for molecules community., Comment: 8 pages, 4 figures. ICML 2024 AI for science workshop (https://openreview.net/forum?id=V4aG4wsoIt). Code at: https://github.com/AustinT/basic-mol-bo-workshop2024

Antibodies, crucial for immune defense, primarily rely on complementarity-determining regions (CDRs) to bind and neutralize antigens, such as viruses. The design of these CDRs determines the antibody's affinity and specificity towards its target. Generative models, particularly denoising diffusion probabilistic models (DDPMs), have shown potential to advance the structure-based design of CDR regions. However, only a limited dataset of bound antibody-antigen structures is available, and generalization to out-of-distribution interfaces remains a challenge. Physics based force-fields, which approximate atomic interactions, offer a coarse but universal source of information to better mold designs to target interfaces. Integrating this foundational information into diffusion models is, therefore, highly desirable. Here, we propose a novel approach to enhance the sampling process of diffusion models by integrating force field energy-based feedback. Our model, DiffForce, employs forces to guide the diffusion sampling process, effectively blending the two distributions. Through extensive experiments, we demonstrate that our method guides the model to sample CDRs with lower energy, enhancing both the structure and sequence of the generated antibodies.

Scaling hyperparameter optimisation to very large datasets remains an open problem in the Gaussian process community. This paper focuses on iterative methods, which use linear system solvers, like conjugate gradients, alternating projections or stochastic gradient descent, to construct an estimate of the marginal likelihood gradient. We discuss three key improvements which are applicable across solvers: (i) a pathwise gradient estimator, which reduces the required number of solver iterations and amortises the computational cost of making predictions, (ii) warm starting linear system solvers with the solution from the previous step, which leads to faster solver convergence at the cost of negligible bias, (iii) early stopping linear system solvers after a limited computational budget, which synergises with warm starting, allowing solver progress to accumulate over multiple marginal likelihood steps. These techniques provide speed-ups of up to $72\times$ when solving to tolerance, and decrease the average residual norm by up to $7\times$ when stopping early., Comment: Advances in Neural Information Processing Systems 2024

Gaussian processes are a versatile probabilistic machine learning model whose effectiveness often depends on good hyperparameters, which are typically learned by maximising the marginal likelihood. In this work, we consider iterative methods, which use iterative linear system solvers to approximate marginal likelihood gradients up to a specified numerical precision, allowing a trade-off between compute time and accuracy of a solution. We introduce a three-level hierarchy of marginal likelihood optimisation for iterative Gaussian processes, and identify that the computational costs are dominated by solving sequential batches of large positive-definite systems of linear equations. We then propose to amortise computations by reusing solutions of linear system solvers as initialisations in the next step, providing a $\textit{warm start}$. Finally, we discuss the necessary conditions and quantify the consequences of warm starts and demonstrate their effectiveness on regression tasks, where warm starts achieve the same results as the conventional procedure while providing up to a $16 \times$ average speed-up among datasets., Comment: Advances in Approximate Bayesian Inference 2024

Relative entropy coding (REC) algorithms encode a random sample following a target distribution $Q$, using a coding distribution $P$ shared between the sender and receiver. Sadly, general REC algorithms suffer from prohibitive encoding times, at least on the order of $2^{D_{\text{KL}}[Q||P]}$, and faster algorithms are limited to very specific settings. This work addresses this issue by introducing a REC scheme utilizing space partitioning to reduce runtime in practical scenarios. We provide theoretical analyses of our method and demonstrate its effectiveness with both toy examples and practical applications. Notably, our method successfully handles REC tasks with $D_{\text{KL}}[Q||P]$ about three times greater than what previous methods can manage, and reduces the bitrate by approximately 5-15% in VAE-based lossless compression on MNIST and INR-based lossy compression on CIFAR-10, compared to previous methods, significantly improving the practicality of REC for neural compression., Comment: NeurIPS 2024 camera-ready version. 31 pages, 12 figures, 5 algorithms

The rich information of electron energy-loss spectroscopy (EELS) comes from the complex inelastic scattering process whereby fast electrons transfer energy and momentum to atoms, exciting bound electrons from their ground states to higher unoccupied states. To quantify EELS, the common practice is to compare the cross-sections integrated within an energy window or fit the observed spectrum with theoretical differential cross-sections calculated from a generalized oscillator strength (GOS) database with experimental parameters. The previous Hartree-Fock-based and DFT-based GOS are calculated from Schr\"odinger's solution of atomic orbitals, which does not include the full relativistic effects. Here, we attempt to go beyond the limitations of the Schr\"odinger solution in the GOS tabulation by including the full relativistic effects using the Dirac equation within the local density approximation, which is particularly important for core-shell electrons of heavy elements with strong spin-orbit coupling. This has been done for all elements in the periodic table (up to Z = 118) for all possible excitation edges using modern computing capabilities and parallelization algorithms. The relativistic effects of fast incoming electrons were included to calculate cross-sections that are specific to the acceleration voltage. We make these tabulated GOS available under an open-source license to the benefit of both academic users as well as allowing integration into commercial solutions., Comment: 52 pages, 8 figures

The evolution of a thematic area undergoes various changes of perspective and adopts new theoretical approaches that arise from the interactions of the community and a wide range of social needs. The advent of digital technologies, such as social networks, underlines this factor by spreading knowledge and forging links between different communities. Web intelligence is now on the verge of raising questions that broaden the understanding of how artificial intelligence impacts the Web of People, Data, and Things, among other factors. To the best of our knowledge, there is no study that has conducted a longitudinal analysis of the evolution of this community. Thus, we investigate in this paper how Web intelligence has evolved in the last twenty years by carrying out a literature review and bibliometric analysis. Concerning the impact of this research study, increasing attention is devoted to determining which are the most influential papers in the community by referring to citation networks and discovering the most popular and pressing topics through a co-citation analysis and the keywords co-occurrence. The results obtained can guide the direction of new research projects in the area and update the scope and places of interest found in current trends and the relevant journals.

Despite substantial progress in machine learning for scientific discovery in recent years, truly de novo design of small molecules which exhibit a property of interest remains a significant challenge. We introduce LambdaZero, a generative active learning approach to search for synthesizable molecules. Powered by deep reinforcement learning, LambdaZero learns to search over the vast space of molecules to discover candidates with a desired property. We apply LambdaZero with molecular docking to design novel small molecules that inhibit the enzyme soluble Epoxide Hydrolase 2 (sEH), while enforcing constraints on synthesizability and drug-likeliness. LambdaZero provides an exponential speedup in terms of the number of calls to the expensive molecular docking oracle, and LambdaZero de novo designed molecules reach docking scores that would otherwise require the virtual screening of a hundred billion molecules. Importantly, LambdaZero discovers novel scaffolds of synthesizable, drug-like inhibitors for sEH. In in vitro experimental validation, a series of ligands from a generated quinazoline-based scaffold were synthesized, and the lead inhibitor N-(4,6-di(pyrrolidin-1-yl)quinazolin-2-yl)-N-methylbenzamide (UM0152893) displayed sub-micromolar enzyme inhibition of sEH.

Swarm robotics promises adaptability to unknown situations and robustness against failures. However, it still struggles with global tasks that require understanding the broader context in which the robots operate, such as identifying the shape of the arena in which the robots are embedded. Biological swarms, such as shoals of fish, flocks of birds, and colonies of insects, routinely solve global geometrical problems through the diffusion of local cues. This paradigm can be explicitly described by mathematical models that could be directly computed and exploited by a robotic swarm. Diffusion over a domain is mathematically encapsulated by the Laplacian, a linear operator that measures the local curvature of a function. Crucially the geometry of a domain can generally be reconstructed from the eigenspectrum of its Laplacian. Here we introduce spectral swarm robotics where robots diffuse information to their neighbors to emulate the Laplacian operator - enabling them to "hear" the spectrum of their arena. We reveal a universal scaling that links the optimal number of robots (a global parameter) with their optimal radius of interaction (a local parameter). We validate experimentally spectral swarm robotics under challenging conditions with the one-shot classification of arena shapes using a sparse swarm of Kilobots. Spectral methods can assist with challenging tasks where robots need to build an emergent consensus on their environment, such as adaptation to unknown terrains, division of labor, or quorum sensing. Spectral methods may extend beyond robotics to analyze and coordinate swarms of agents of various natures, such as traffic or crowds, and to better understand the long-range dynamics of natural systems emerging from short-range interactions.

Correctly capturing the symmetry transformations of data can lead to efficient models with strong generalization capabilities, though methods incorporating symmetries often require prior knowledge. While recent advancements have been made in learning those symmetries directly from the dataset, most of this work has focused on the discriminative setting. In this paper, we take inspiration from group theoretic ideas to construct a generative model that explicitly aims to capture the data's approximate symmetries. This results in a model that, given a prespecified broad set of possible symmetries, learns to what extent, if at all, those symmetries are actually present. Our model can be seen as a generative process for data augmentation. We provide a simple algorithm for learning our generative model and empirically demonstrate its ability to capture symmetries under affine and color transformations, in an interpretable way. Combining our symmetry model with standard generative models results in higher marginal test-log-likelihoods and improved data efficiency.

In this paper, a novel approach to define the optimal bidding of renewable-only virtual power plants (RVPPs) in the day-ahead, secondary reserve, and intra-day markets is proposed. To this aim, a robust optimization algorithm is developed to account for the asymmetric nature of the uncertainties that characterize the market prices, as well as the energy production of the RVPP stochastic sources and flexible demand consumption. Simulation results show increased RVPP benefits compared to other existing solutions and demonstrate the potential of renewable sources to further increase their economic competitiveness. The simplicity of the implementation, the computational efficiency, and the flexible robustness are also verified.

We investigate the problem of explainability for machine learning models, focusing on Feature Attribution Methods (FAMs) that evaluate feature importance through perturbation tests. Despite their utility, FAMs struggle to distinguish the contributions of different features, when their prediction changes are similar after perturbation. To enhance FAMs' discriminative power, we introduce Feature Attribution with Necessity and Sufficiency (FANS), which find a neighborhood of the input such that perturbing samples within this neighborhood have a high Probability of being Necessity and Sufficiency (PNS) cause for the change in predictions, and use this PNS as the importance of the feature. Specifically, FANS compute this PNS via a heuristic strategy for estimating the neighborhood and a perturbation test involving two stages (factual and interventional) for counterfactual reasoning. To generate counterfactual samples, we use a resampling-based approach on the observed samples to approximate the required conditional distribution. We demonstrate that FANS outperforms existing attribution methods on six benchmarks. Please refer to the source code via \url{https://github.com/DMIRLAB-Group/FANS}., Comment: Accepted in the Proceedings of the 41st International Conference on Machine Learning (ICML2024)

Speech technologies rely on capturing a speaker's voice variability while obtaining comprehensive language information. Textual prompts and sentence selection methods have been proposed in the literature to comprise such adequate phonetic data, referred to as a phonetically rich \textit{corpus}. However, they are still insufficient for acoustic modeling, especially critical for languages with limited resources. Hence, this paper proposes a novel approach and outlines the methodological aspects required to create a \textit{corpus} with broad phonetic coverage for a low-resourced language, Brazilian Portuguese. Our methodology includes text dataset collection up to a sentence selection algorithm based on triphone distribution. Furthermore, we propose a new phonemic classification according to acoustic-articulatory speech features since the absolute number of distinct triphones, or low-probability triphones, does not guarantee an adequate representation of every possible combination. Using our algorithm, we achieve a 55.8\% higher percentage of distinct triphones -- for samples of similar size -- while the currently available phonetic-rich corpus, CETUC and TTS-Portuguese, 12.6\% and 12.3\% in comparison to a non-phonetically rich dataset.

The inadequate mixing of conventional Markov Chain Monte Carlo (MCMC) methods for multi-modal distributions presents a significant challenge in practical applications such as Bayesian inference and molecular dynamics. Addressing this, we propose Diffusive Gibbs Sampling (DiGS), an innovative family of sampling methods designed for effective sampling from distributions characterized by distant and disconnected modes. DiGS integrates recent developments in diffusion models, leveraging Gaussian convolution to create an auxiliary noisy distribution that bridges isolated modes in the original space and applying Gibbs sampling to alternately draw samples from both spaces. A novel Metropolis-within-Gibbs scheme is proposed to enhance mixing in the denoising sampling step. DiGS exhibits a better mixing property for sampling multi-modal distributions than state-of-the-art methods such as parallel tempering, attaining substantially improved performance across various tasks, including mixtures of Gaussians, Bayesian neural networks and molecular dynamics., Comment: Accepted for publication at ICML 2024. Code available: https://github.com/Wenlin-Chen/DiGS

We generalize, imposing the field equations only at dominant order, the Isaacson formula for the gravitational wave (GW) energy-momentum tensor (EMT) to the class of Horndeski theories in which the tensor modes travel at the speed of light (reduced Horndeski theories) and scalar waves are present. We discuss important particular cases such as: theories where scalar waves are also luminal and theories in which the transverse-traceless gauge can be achieved in an arbitrary open set. The vanishing of the trace of the gravitational wave energy-momentum tensor is obtained for theories in which all wave perturbations propagate at the speed of light. The trace is shown not to vanish trivially in other cases. We obtain, as a particular case of our general result, the GW EMTs, in a Brans-Dicke theory, both in the Einstein frame, recovering previous results in the literature, and in the Jordan frame, thereby showing the GW EMT is not conformally invariant. We further prove that there exists a subclass of reduced Horndeski theories where, in contrast to general relativity, the divergence of the GW EMT does not vanish even after the imposition of the full equations of motion, assuming an eikonal solution.

White light spectral interferometry is applied to measure the refractive index in absorbing liquids in the spectral range from 400 to 1000 nm. We analyze the influence of absorption on the visibility of interferometric fringes and, accordingly, on the measurement of the refractive index. Further, we show that the refractive index in the absorption band can be retrieved by a two step process. The procedure requires the use of two samples of different thickness, the thicker one to retrieve the refractive index in the transparent region and the thinnest to obtain the data in the absorption region. First, the refractive index values are retrieved with good accuracy in the transparent region of the material for 1 mm thick samples. Second, these refractive index values serve also to precisely calculate the thickness of a thinner sample (about 150 micrometers) since the accuracy of the method depends strongly on the thickness of the sample. Finally, the refractive index is recovered for the entire spectral range. Copyright 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement. See corresponding link in the pdf file., Comment: 9 pages, 6 figures, 2 tables

In the current landscape of deep learning research, there is a predominant emphasis on achieving high predictive accuracy in supervised tasks involving large image and language datasets. However, a broader perspective reveals a multitude of overlooked metrics, tasks, and data types, such as uncertainty, active and continual learning, and scientific data, that demand attention. Bayesian deep learning (BDL) constitutes a promising avenue, offering advantages across these diverse settings. This paper posits that BDL can elevate the capabilities of deep learning. It revisits the strengths of BDL, acknowledges existing challenges, and highlights some exciting research avenues aimed at addressing these obstacles. Looking ahead, the discussion focuses on possible ways to combine large-scale foundation models with BDL to unlock their full potential., Comment: Proceedings of the 41st International Conference on Machine Learning, Vienna, Austria. PMLR 235, 2024

Couplings between biochemical and mechanical processes have a profound impact on embryonic development. However, in-vitro studies capable of quantifying these interactions have remained elusive. Here, we investigate a synthetic system where a DNA reaction-diffusion (RD) front is advected by a turbulent flow generated by active matter (AM) flows in a quasi-one-dimensional geometry. Whereas the dynamics of simple RD fronts solely depend on the reaction and diffusion rates, we show that RD-AM front propagation is also influenced by the confinement geometry. We first experimentally dissected the different components of the reaction-diffusion-advection process by knocking out reaction or advection and observed how RD-AM allows for faster transport over large distances, avoiding dilution. We then show how confinement impacts active matter flow: while changes in instantaneous flow velocities are small; correlation times are dramatically increased with decreasing confinement. As a result, RD-AM front speed increased 3 to 9-fold compared to a RD one. This RD-AM experimental model provides a framework for the rational engineering of complex spatiotemporal processes observed in living systems. It will reinforce our understanding of how macro-scale patterns and structures emerge from microscopic components in non-equilibrium systems.

This paper presents a data-driven methodology for estimating Under Frequency Load Shedding (UFLS) in small power systems. UFLS plays a vital role in maintaining system stability by shedding load when the frequency drops below a specified threshold following loss of generation. Using a dynamic System Frequency Response (SFR) model we generate different values of UFLS (i.e., labels) predicated on a set of carefully selected operating conditions (i.e., features). Machine Learning (ML) algorithms are then applied to learn the relationship between chosen features and the UFLS labels. A novel regression tree and the Tobit model are suggested for this purpose and we show how the resulting non-linear model can be directly incorporated into a Mixed Integer Linear Programming (MILP) problem. The trained model can be used to estimate UFLS in security-constrained operational planning problems, improving frequency response, optimizing reserve allocation, and reducing costs. The methodology is applied to the La Palma island power system, demonstrating its accuracy and effectiveness. The results confirm that the amount of UFLS can be estimated with the Mean Absolute Error (MAE) as small as 0.213 MW for the whole process, with a model that is representable as a MILP for use in scheduling problems such as unit commitment among others.

The conventional approach to solving the unit commitment problem involves discrete intervals at an hourly scale, particularly when integrating frequency dynamics to formulate a frequency-constrained unit commitment. To overcome this limitation, a novel continuous-time frequency-constrained unit commitment framework is proposed in this paper. In this approach, Bernstein polynomials represent continuous variables in the unit commitment problem and enable the calculation of frequency response-related metrics such as the rate of change of frequency, quasi-steady-state frequency, and frequency nadir. Notably, startup and shut-down trajectories are meticulously considered, transforming the formulation into a fully continuous-time model and simplifying constraints related to variable continuity. To address the complexities associated with integrating the obtained non-linear frequency nadir constraint into a mixed-integer linear problem, an alternative data-driven frequency nadir constraint is proposed, which accurately constrains frequency nadir deviations throughout the time interval. To validate the proposed model, it is applied to the real-life network of the Spanish Island of La Palma. The results demonstrate the effectiveness of the proposed formulation, indicating that the model is solved timely while mitigating the impact of intra-hour real-time power fluctuations on system frequency.