Energy efficiency is a crucial requirement for enabling powerful artificial intelligence applications at the microedge. Hardware acceleration with frugal architectural allocation is an effective method for reducing energy. Many emerging applications also require the systems design to incorporate interpretable decision models to establish responsibility and transparency. The design needs to provision for additional resources to provide reachable states in real-world data scenarios, defining conflicting design tradeoffs between energy efficiency. is challenging. Recently a new machine learning algorithm, called the Tsetlin machine, has been proposed. The algorithm is fundamentally based on the principles of finite-state automata and benefits from natural logic underpinning rather than arithmetic. In this paper, we investigate methods of energy-frugal artificial intelligence hardware design by suitably tuning the hyperparameters, while maintaining high learning efficacy. To demonstrate interpretability, we use reachability and game-theoretic analysis in two simulation environments: a SystemC model to study the bounded state transitions in the presence of hardware faults and Nash equilibrium between states to analyze the learning convergence. Our analyses provides the first insights into conflicting design tradeoffs involved in energy-efficient and interpretable decision models for this new artificial intelligence hardware architecture. We show that frugal resource allocation coupled with systematic prodigality between randomized reinforcements can provide decisive energy reduction while also achieving robust and interpretable learning.

However, state-of-the-art deep SNNs (including Spikformer and SEW ResNet) suffer from non-spike computations (integer-float multiplications) caused by the structure of their residual connection. These non-spike computations increase SNNs' power consumption and make them unsuitable for deployment on mainstream neuromorphic hardware, which only supports spike operations. In this paper, we propose a hardware-friendly spike-driven residual learning architecture for SNNs to avoid non-spike computations. Based on this residual design, we develop Spikingformer, a pure transformer-based spiking neural network. We evaluate Spikingformer on ImageNet, CIFAR10, CIFAR100, CIFAR10-DVS and DVS128 Gesture datasets, and demonstrate that Spikingformer outperforms the state-of-the-art in directly trained pure SNNs as a novel advanced backbone (75.85% top-1 accuracy on ImageNet, + 1.04% compared with Spikformer). Furthermore, our experiments verify that Spikingformer effectively avoids non-spike computations and significantly reduces energy consumption by 57.34% compared with Spikformer on ImageNet. To our best knowledge, this is the first time that a pure event-driven transformer-based SNN has been developed. Codes will be available at Spikingformer.

Real-time multi-task multi-model (MTMM) workloads, a new form of deep learning inference workloads, are emerging for applications areas like extended reality (XR) to support metaverse use cases. These workloads combine user interactivity with computationally complex machine learning (ML) activities. Compared to standard ML applications, these ML workloads present unique difficulties and constraints. Real-time MTMM workloads impose heterogeneity and concurrency requirements on future ML systems and devices, necessitating the development of new capabilities. This paper begins with a discussion of the various characteristics of these real-time MTMM ML workloads and presents an ontology for evaluating the performance of future ML hardware for XR systems. Next, we present XRBENCH, a collection of MTMM ML tasks, models, and usage scenarios that execute these models in three representative ways: cascaded, concurrent, and cascaded-concurrent for XR use cases. Finally, we emphasize the need for new metrics that capture the requirements properly. We hope that our work will stimulate research and lead to the development of a new generation of ML systems for XR use cases. XRBench is available as an open-source project: https://github.com/XRBench

Synthesis prediction is a key accelerator for the rapid design of advanced materials. However, determining synthesis variables such as the choice of precursor materials is challenging for inorganic materials because the sequence of reactions during heating is not well understood. In this work, we use a knowledge base of 29,900 solid-state synthesis recipes, text-mined from the scientific literature, to automatically learn which precursors to recommend for the synthesis of a novel target material. The data-driven approach learns chemical similarity of materials and refers the synthesis of a new target to precedent synthesis procedures of similar materials, mimicking human synthesis design. When proposing five precursor sets for each of 2,654 unseen test target materials, the recommendation strategy achieves a success rate of at least 82%. Our approach captures decades of heuristic synthesis data in a mathematical form, making it accessible for use in recommendation engines and autonomous laboratories.

Gas source localization (GSL) with an autonomous robot is a problem with many prospective applications, from finding pipe leaks to emergency-response scenarios. In this work, we present a new method to perform GSL in realistic indoor environments, featuring obstacles and turbulent flow. Given the highly complex relationship between the source position and the measurements available to the robot (the single-point gas concentration, and the wind vector) we propose an observation model that derives from contrasting the online, real-time simulation of the gas dispersion from any candidate source localization against a gas concentration map built from sensor readings. To account for a convenient and grounded integration of both into a probabilistic estimation framework, we introduce the concept of probabilistic gas-hit maps, which provide a higher level of abstraction to model the time-dependent nature of gas dispersion. Results from both simulated and real experiments show the capabilities of our current proposal to deal with source localization in complex indoor environments.

A variety of established methods exist for modeling data, ranging from traditional machine learning techniques (linear regression, ridge regression, polynomial regression) to deep learning approaches (neural networks). However, these methods suffer from constraints and/or interpretability, such as limiting the model to a particular shape (e.g., linear), or being too complex to interpret (black box models). Symbolic Regression (SR) is less constrained and searches through the mathematical space of equations. SR allows for discovering a broader range of functional relationships, including those with nonlinear or intricate interactions between variables.

Abstract-- In this paper we consider the problem of allowing a humanoid robot that is subject to a persistent disturbance, in the form of a payload-carrying task, to follow given planned footsteps. MPC is augmented with terms handling the disturbance and regularizing the parameter. Finally, the effect of using the parametrization on the computational time of the controller is briefly studied. The high-level control layer typically utilizes "template" models to reason about the center of mass and feet trajectories [2], while the whole-body control layer uses the robot full model to track the adapted trajectories (see Figure 1). This paper focuses on designing a high-level trajectory adjustment controller leveraging a Figure 1: The controller highlighted in a typical multi-layer template model to allow for humanoid robots locomotion bipedal locomotion control architecture.

It has been found that certain biological organisms, such as Erodium seeds and Scincus scincus, are capable of effectively and efficiently burying themselves in soil. Biological Organisms employ various locomotion modes, including coiling and uncoiling motions, asymmetric body twisting, and undulating movements that generate motion waves. The coiling-uncoiling motion drives a seed awn to bury itself like a corkscrew, while sandfish skinks use undulatory swimming, which can be thought of as a 2D version of helical motion. Studying burrowing behavior aims to understand how animals navigate underground, whether in their natural burrows or underground habitats, and to implement this knowledge in solving geotechnical penetration problems. Underground horizontal burrowing is challenging due to overcoming the resistance of interaction forces of granular media to move forward. Inspired by the burrowing behavior of seed-awn and sandfish skink, a horizontal self-burrowing robot is developed. The robot is driven by two augers and stabilized by a fin structure. The robot's burrowing behavior is studied in a laboratory setting. It is found that rotation and propulsive motion along the axis of the auger's helical shape significantly reduce granular media's resistance against horizontal penetration by breaking kinematic symmetry or granular media boundary. Additional thrusting and dragging tests were performed to examine the propulsive and resistive forces and unify the observed burrowing behaviors. The tests revealed that the rotation of an auger not only reduces the resistive force and generates a propulsive force, which is influenced by the auger geometry, rotational speed, and direction. As a result, the burrowing behavior of the robot can be predicted using the geometry-rotation-force relations.

A parametric adaptive physics-informed greedy Latent Space Dynamics Identification (gLaSDI) method is proposed for accurate, efficient, and robust data-driven reduced-order modeling of high-dimensional nonlinear dynamical systems. In the proposed gLaSDI framework, an autoencoder discovers intrinsic nonlinear latent representations of high-dimensional data, while dynamics identification (DI) models capture local latent-space dynamics. An interactive training algorithm is adopted for the autoencoder and local DI models, which enables identification of simple latent-space dynamics and enhances accuracy and efficiency of data-driven reduced-order modeling. To maximize and accelerate the exploration of the parameter space for the optimal model performance, an adaptive greedy sampling algorithm integrated with a physics-informed residual-based error indicator and random-subset evaluation is introduced to search for the optimal training samples on the fly. Further, to exploit local latent-space dynamics captured by the local DI models for an improved modeling accuracy with a minimum number of local DI models in the parameter space, a k-nearest neighbor convex interpolation scheme is employed. The effectiveness of the proposed framework is demonstrated by modeling various nonlinear dynamical problems, including Burgers equations, nonlinear heat conduction, and radial advection. The proposed adaptive greedy sampling outperforms the conventional predefined uniform sampling in terms of accuracy. Compared with the high-fidelity models, gLaSDI achieves 17 to 2,658x speed-up with 1 to 5% relative errors.

In this paper, we propose a novel algorithm for energy-efficient, low-latency, accurate inference at the wireless edge, in the context of 6G networks endowed with reconfigurable intelligent surfaces (RISs). We consider a scenario where new data are continuously generated/collected by a set of devices and are handled through a dynamic queueing system. Building on the marriage between Lyapunov stochastic optimization and deep reinforcement learning (DRL), we devise a dynamic learning algorithm that jointly optimizes the data compression scheme, the allocation of radio resources (i.e., power, transmission precoding), the computation resources (i.e., CPU cycles), and the RIS reflectivity parameters (i.e., phase shifts), with the aim of performing energy-efficient edge classification with end-to-end (E2E) delay and inference accuracy constraints. The proposed strategy enables dynamic control of the system and of the wireless propagation environment, performing a low-complexity optimization on a per-slot basis while dealing with time-varying radio channels and task arrivals, whose statistics are unknown. Numerical results assess the performance of the proposed RIS-empowered edge inference strategy in terms of trade-off between energy, delay, and accuracy of a classification task.