Scaling model capacity enhances its capabilities but significantly increases computation. Mixture-of-Experts models (MoEs) address this by allowing model capacity to scale without substantially increasing training or inference costs. Despite their promising results, MoE models encounter several challenges. Primarily, the dispersion of training tokens across multiple experts can lead to underfitting, particularly for infrequent tokens. Additionally, while fixed routing mechanisms can mitigate this issue, they compromise on the diversity of representations. In this paper, we propose MaskMoE, a method designed to enhance token-level learning by employing a routing masking technique within the Mixture-of-Experts model. MaskMoE is capable of maintaining representation diversity while achieving more comprehensive training. Experimental results demonstrate that our method outperforms previous dominant Mixture-of-Experts models in both perplexity (PPL) and downstream tasks.

The paper reflects on the future role of AI in scientific research, with a special focus on turbulence studies, and examines the evolution of AI, particularly through Diffusion Models rooted in non-equilibrium statistical mechanics. It underscores the significant impact of AI on advancing reduced, Lagrangian models of turbulence through innovative use of deep neural networks. Additionally, the paper reviews various other AI applications in turbulence research and outlines potential challenges and opportunities in the concurrent advancement of AI and statistical hydrodynamics. This discussion sets the stage for a future where AI and turbulence research are intricately intertwined, leading to more profound insights and advancements in both fields.

Bridging the gap between classical scientific computing and machine learning, the emerging field called scientific machine learning has introduced a completely different framework to compute the solutions of partial differential equations (PDEs). The forefront of this evolution lies within the area of physicsinformed neural networks (PINNs) [18, 10]. Based on the universal approximation property of deep neural networks, the PINNs approximate the solutions of PDEs by incorporating underlying physics into deep neural networks. Due to their simplicity and flexibility in handling a wide range of physical problems involving PDEs, PINNs have recently gained great attention and have been applied to various fields in computational science: bio-medical science [22, 11], fluids mechanics [20, 19, 9, 4, 21], uncertainty quantification [30, 32, 29] and meta-material design [12, 3]. Moreover, since the PINNs utilize randomlyselected collocation points as training samples in the spatio-temporal domain, the PINNs are available for high-dimensional PDEs [24, 6], on the domains with complex geometries [13, 23, 26, 14]. However, despite the significant empirical success of PINNs, we only have limited knowledge about the behavior of these constrained neural networks during their training process, and the training of PINNs often fails. In particular, since neural networks typically assume a smooth prior, it is often challenging to train PINNs to learn a solution with a sharp transition which poses significant obstacles for the model prediction. For example, some recent studies have illustrated that training PINNs with fully-connected neural networks usually suffers from so-called spectral bias or F-principle meaning that it is difficult for PINNs to learn functions with high frequencies [17, 2, 25].

High-performance scientific simulations, important for comprehension of complex systems, encounter computational challenges especially when exploring extensive parameter spaces. There has been an increasing interest in developing deep neural networks (DNNs) as surrogate models capable of accelerating the simulations. However, existing approaches for training these DNN surrogates rely on extensive simulation data which are heuristically selected and generated with expensive computation -- a challenge under-explored in the literature. In this paper, we investigate the potential of incorporating active learning into DNN surrogate training. This allows intelligent and objective selection of training simulations, reducing the need to generate extensive simulation data as well as the dependency of the performance of DNN surrogates on pre-defined training simulations. In the problem context of constructing DNN surrogates for diffusion equations with sources, we examine the efficacy of diversity- and uncertainty-based strategies for selecting training simulations, considering two different DNN architecture. The results set the groundwork for developing the high-performance computing infrastructure for Smart Surrogates that supports on-the-fly generation of simulation data steered by active learning strategies to potentially improve the efficiency of scientific simulations.

Volumetric video based on Neural Radiance Field (NeRF) holds vast potential for various 3D applications, but its substantial data volume poses significant challenges for compression and transmission. Current NeRF compression lacks the flexibility to adjust video quality and bitrate within a single model for various network and device capacities. To address these issues, we propose HPC, a novel hierarchical progressive volumetric video coding framework achieving variable bitrate using a single model. Specifically, HPC introduces a hierarchical representation with a multi-resolution residual radiance field to reduce temporal redundancy in long-duration sequences while simultaneously generating various levels of detail. Then, we propose an end-to-end progressive learning approach with a multi-rate-distortion loss function to jointly optimize both hierarchical representation and compression. Our HPC trained only once can realize multiple compression levels, while the current methods need to train multiple fixed-bitrate models for different rate-distortion (RD) tradeoffs. Extensive experiments demonstrate that HPC achieves flexible quality levels with variable bitrate by a single model and exhibits competitive RD performance, even outperforming fixed-bitrate models across various datasets.

Multi-agent systems (MAS) constitute a significant role in exploring machine intelligence and advanced applications. In order to deeply investigate complicated interactions within MAS scenarios, we originally propose "GNN for MBRL" model, which utilizes a state-spaced Graph Neural Networks with Model-based Reinforcement Learning to address specific MAS missions (e.g., Billiard-Avoidance, Autonomous Driving Cars). In detail, we firstly used GNN model to predict future states and trajectories of multiple agents, then applied the Cross-Entropy Method (CEM) optimized Model Predictive Control to assist the ego-agent planning actions and successfully accomplish certain MAS tasks.

In this work, we propose a novel method for Bayesian Networks (BNs) structure elicitation that is based on the initialization of several LLMs with different experiences, independently querying them to create a structure of the BN, and further obtaining the final structure by majority voting. We compare the method with one alternative method on various widely and not widely known BNs of different sizes and study the scalability of both methods on them. We also propose an approach to check the contamination of BNs in LLM, which shows that some widely known BNs are inapplicable for testing the LLM usage for BNs structure elicitation. We also show that some BNs may be inapplicable for such experiments because their node names are indistinguishable. The experiments on the other BNs show that our method performs better than the existing method with one of the three studied LLMs; however, the performance of both methods significantly decreases with the increase in BN size.

ABSTRACT We present a flow-based generative approach to emulate grids of stellar evolutionary models. By interpreting the input parameters and output properties of these models as multi-dimensional probability distributions, we train conditional normalizing flows to learn and predict the complex relationships between grid inputs and outputs in the form of conditional joint distributions. Leveraging the expressive power and versatility of these flows, we showcase their ability to emulate a variety of evolutionary tracks and isochrones across a continuous range of input parameters. In addition, we describe a simple Bayesian approach for estimating stellar parameters using these flows and demonstrate its application to asteroseismic datasets of red giants observed by the Kepler mission. By applying this approach to red giants in open clusters NGC 6791 and NGC 6819, we illustrate how large age uncertainties can arise when fitting only to global asteroseismic and spectroscopic parameters without prior information on initial helium abundances and mixing length parameter values. We also conduct inference using the flow at a large scale by determining revised estimates of masses and radii for 15,388 field red giants. These estimates show improved agreement with results from existing grid-based modelling, reveal distinct population-level features in the red clump, and suggest that the masses of Kepler red giants previously determined using the corrected asteroseismic scaling relations have been overestimated by 5 10%.

Mergers of binary neutron stars (BNSs) emit signals in both the gravitational-wave (GW) and electromagnetic (EM) spectra. Famously, the 2017 multi-messenger observation of GW170817 led to scientific discoveries across cosmology, nuclear physics, and gravity. Central to these results were the sky localization and distance obtained from GW data, which, in the case of GW170817, helped to identify the associated EM transient, AT 2017gfo, 11 hours after the GW signal. Fast analysis of GW data is critical for directing time-sensitive EM observations; however, due to challenges arising from the length and complexity of signals, it is often necessary to make approximations that sacrifice accuracy. Here, we develop a machine learning approach that performs complete BNS inference in just one second without making any such approximations. This is enabled by a new method for explicit integration of physical domain knowledge into neural networks. Our approach enhances multi-messenger observations by providing (i) accurate localization even before the merger; (ii) improved localization precision by $\sim30\%$ compared to approximate low-latency methods; and (iii) detailed information on luminosity distance, inclination, and masses, which can be used to prioritize expensive telescope time. Additionally, the flexibility and reduced cost of our method open new opportunities for equation-of-state and waveform systematics studies. Finally, we demonstrate that our method scales to extremely long signals, up to an hour in length, thus serving as a blueprint for data analysis for next-generation ground- and space-based detectors.

This paper discusses developments for a multi-limb morphogenetic UAV, MorphoGear, that is capable of both aerial flight and ground locomotion. A hybrid path planning algorithm based on A* strategy has been developed enabling seamless transition between air-to-ground navigation modes, thereby enhancing robot's mobility in complex environments. Moreover, precise path following is achieved during ground locomotion with a Model Predictive Control (MPC) architecture for its novel walking behaviour. Experimental validation was conducted in the Unity simulation environment utilizing Python scripts to compute control values. The algorithms' performance is validated by the Root Mean Squared Error (RMSE) of 0.91 cm and a maximum error of 1.85 cm, as demonstrated by the results. These developments highlight the adaptability of MorphoGear in navigation through cluttered environments, establishing it as a usable tool in autonomous exploration, both aerial and ground-based.