Artificial neural networks (ANNs) have permeated various disciplinary domains, ranging from bioinformatics to financial analytics, where their application has become an indispensable facet of contemporary scientific research endeavors. However, the inherent limitations of traditional neural networks arise due to their relatively fixed network structures and activation functions. 1, The type of activation function is single and relatively fixed, which leads to poor "unit representation ability" of the network, and it is often used to solve simple problems with very complex networks; 2, the network structure is not adaptive, it is easy to cause network structure redundant or insufficient. To address the aforementioned issues, this study proposes a novel neural network called X-Net. By utilizing our designed Alternating Backpropagation mechanism, X-Net dynamically selects appropriate activation functions based on derivative information during training to enhance the network's representation capability for specific tasks. Simultaneously, it accurately adjusts the network structure at the neuron level to accommodate tasks of varying complexities and reduce computational costs. Through a series of experiments, we demonstrate the dual advantages of X-Net in terms of reducing model size and improving representation power. Specifically, in terms of the number of parameters, X-Net is only 3$\%$ of baselines on average, and only 1.4$\%$ under some tasks. In terms of representation ability, X-Net can achieve an average $R^2$=0.985 on the fitting task by only optimizing the activation function without introducing any parameters. Finally, we also tested the ability of X-Net to help scientific discovery on data from multiple disciplines such as society, energy, environment, and aerospace, and achieved concise and good results.

Reservoir computing (RC) systems can efficiently forecast chaotic time series using nonlinear dynamical properties of an artificial neural network of random connections. The versatility of RC systems has motivated further research on both hardware counterparts of traditional RC algorithms and more efficient RC-like schemes. Inspired by the nonlinear processes in a living biological brain and using solitary waves excited on the surface of a flowing liquid film, in this paper we experimentally validate a physical RC system that substitutes the effect of randomness for a nonlinear transformation of input data. Carrying out all operations using a microcontroller with a minimal computational power, we demonstrate that the so-designed RC system serves as a technically simple hardware counterpart to the `next-generation' improvement of the traditional RC algorithm.

This work examines the effects of variations in machine learning training regimes and learning paradigms on the corresponding energy consumption. While increasing data availability and innovation in high-performance hardware fuels the training of sophisticated models, it also supports the fading perception of energy consumption and carbon emission. Therefore, the goal of this work is to create awareness about the energy impact of general training parameters and processes, from learning rate over batch size to knowledge transfer. Multiple setups with different hyperparameter initializations are evaluated on two different hardware configurations to obtain meaningful results. Experiments on pretraining and multitask training are conducted on top of the baseline results to determine their potential towards sustainable machine learning.

The burgeoning field of Large Language Models (LLMs), exemplified by sophisticated models like OpenAI's ChatGPT, represents a significant advancement in artificial intelligence. These models, however, bring forth substantial challenges in the high consumption of computational, memory, energy, and financial resources, especially in environments with limited resource capabilities. This survey aims to systematically address these challenges by reviewing a broad spectrum of techniques designed to enhance the resource efficiency of LLMs. We categorize methods based on their optimization focus: computational, memory, energy, financial, and network resources and their applicability across various stages of an LLM's lifecycle, including architecture design, pretraining, finetuning, and system design. Additionally, the survey introduces a nuanced categorization of resource efficiency techniques by their specific resource types, which uncovers the intricate relationships and mappings between various resources and corresponding optimization techniques. A standardized set of evaluation metrics and datasets is also presented to facilitate consistent and fair comparisons across different models and techniques. By offering a comprehensive overview of the current sota and identifying open research avenues, this survey serves as a foundational reference for researchers and practitioners, aiding them in developing more sustainable and efficient LLMs in a rapidly evolving landscape.

We propose a fixed-lag smoother-based sensor fusion architecture to leverage the complementary benefits of range-based sensors and visual-inertial odometry (VIO) for localization. We use two fixed-lag smoothers (FLS) to decouple accurate state estimation and high-rate pose generation for closed-loop control. The first FLS combines ultrawideband (UWB)-based range measurements and VIO to estimate the robot trajectory and any systematic biases that affect the range measurements in cluttered environments. The second FLS estimates smooth corrections to VIO to generate pose estimates at a high rate for online control. The proposed method is lightweight and can run on a computationally constrained micro-aerial vehicle (MAV). We validate our approach through closed-loop flight tests involving dynamic trajectories in multiple real-world cluttered indoor environments. Our method achieves decimeter-to-sub-decimeter-level positioning accuracy using off-the-shelf sensors and decimeter-level tracking accuracy with minimally-tuned open-source controllers.

Multi-legged robots with six or more legs are not in common use, despite designs with superior stability, maneuverability, and a low number of actuators being available for over 20 years. This may be in part due to the difficulty in modeling multi-legged motion with slipping and producing reliable predictions of body velocity. Here we present a detailed measurement of the foot contact forces in a hexapedal robot with multiple sliding contacts, and provide an algorithm for predicting these contact forces and the body velocity. The algorithm relies on the recently published observation that even while slipping, multi-legged robots are principally kinematic, and employ a friction law ansatz that allows us to compute the shape-change to body-velocity connection and the foot contact forces. This results in the ability to simulate motion plans for a large number of contacts, each potentially with slipping. Furthermore, in homogeneous environments, this kind of simulation can run in (parallel) logarithmic time of the planning horizon.

Hyperspectral images (HSIs) contain rich spectral and spatial information. Motivated by the success of transformers in the field of natural language processing and computer vision where they have shown the ability to learn long range dependencies within input data, recent research has focused on using transformers for HSIs. However, current state-of-the-art hyperspectral transformers only tokenize the input HSI sample along the spectral dimension, resulting in the under-utilization of spatial information. Moreover, transformers are known to be data-hungry and their performance relies heavily on large-scale pretraining, which is challenging due to limited annotated hyperspectral data. Therefore, the full potential of HSI transformers has not been fully realized. To overcome these limitations, we propose a novel factorized spectral-spatial transformer that incorporates factorized self-supervised pretraining procedures, leading to significant improvements in performance. The factorization of the inputs allows the spectral and spatial transformers to better capture the interactions within the hyperspectral data cubes. Inspired by masked image modeling pretraining, we also devise efficient masking strategies for pretraining each of the spectral and spatial transformers. We conduct experiments on six publicly available datasets for HSI classification task and demonstrate that our model achieves state-of-the-art performance in all the datasets. The code for our model will be made available at https://github.com/csiro-robotics/factoformer.

For situations that may benefit from information sharing among datasets, e.g., population-based SHM of similar structures, the hierarchical Bayesian approach provides a useful modelling structure. Hierarchical Bayesian models learn statistical distributions at the population (or parent) and the domain levels simultaneously, to bolster statistical strength among the parameters. As a result, variance is reduced among the parameter estimates, particularly when data are limited. In this paper, a combined probabilistic FRF model is developed for a small population of nominally-identical helicopter blades, using a hierarchical Bayesian structure, to support information transfer in the context of sparse data. The modelling approach is also demonstrated in a traditional SHM context, for a single helicopter blade exposed to varying temperatures, to show how the inclusion of physics-based knowledge can improve generalisation beyond the training data, in the context of scarce data. These models address critical challenges in SHM, by accommodating benign variations that present as differences in the underlying dynamics, while also considering (and utilising), the similarities among the domains.

Current research in the computer vision field mainly focuses on improving Deep Learning (DL) correctness and inference time performance. However, there is still little work on the huge carbon footprint that has training DL models. This study aims to analyze the impact of the model architecture and training environment when training greener computer vision models. We divide this goal into two research questions. First, we analyze the effects of model architecture on achieving greener models while keeping correctness at optimal levels. Second, we study the influence of the training environment on producing greener models. To investigate these relationships, we collect multiple metrics related to energy efficiency and model correctness during the models' training. Then, we outline the trade-offs between the measured energy efficiency and the models' correctness regarding model architecture, and their relationship with the training environment. We conduct this research in the context of a computer vision system for image classification. In conclusion, we show that selecting the proper model architecture and training environment can reduce energy consumption dramatically (up to 81.38%) at the cost of negligible decreases in correctness. Also, we find evidence that GPUs should scale with the models' computational complexity for better energy efficiency.

The field called physics-informed machine learning combines the strengths of physics-based models and data-driven techniques to achieve a deeper understanding and improved predictive capabilities for complex physical systems [36, 61]. The rapidly growing interest in this interdisciplinary approach is largely motivated by the increasing capabilities of computers to store and process large quantities of data, along with the decreasing costs of sensors and computers that capture and handle data from physical systems. Machine learning for differential equations can broadly be divided into two categories: the forward problem, which involves predicting future states from an initial state, and the inverse problem, which entails learning a system or parts of it from data. A wealth of recent literature exists on machine learning for the forward problem in the context of partial differential equations (PDEs). The proposed methods include neural-network-based substitutes for numerical solvers [20, 21, 58, 53], but also methods that can aid the solution process, e.g. by optimizing the discretization to be used in a solver [2]. The focus of this paper is on the inverse problem, and much of the foundation for our proposed model can be found in recent advances in learning neural network models for ordinary differential equations (ODEs). Specifically, we build on recent works on models that incorporate Hamiltonian mechanics and related structures that underlie the physical systems we seek to model.