Search

The estimation and management of motor temperature are important for the continuous movements of robots. In this study, we propose an online learning method of thermal model parameters of motors for an accurate estimation of motor core temperature. Also, we propose a management method of motor core temperature using the updated model and anomaly detection method of motors. Finally, we apply this method to the muscles of the musculoskeletal humanoid and verify the ability of continuous movements., Comment: Accepted at IEEE Robotics and Automation Letters

This paper introduces LLM-jp, a cross-organizational project for the research and development of Japanese large language models (LLMs). LLM-jp aims to develop open-source and strong Japanese LLMs, and as of this writing, more than 1,500 participants from academia and industry are working together for this purpose. This paper presents the background of the establishment of LLM-jp, summaries of its activities, and technical reports on the LLMs developed by LLM-jp. For the latest activities, visit https://llm-jp.nii.ac.jp/en/.

In reinforcement learning (RL) with experience replay, experiences stored in a replay buffer influence the RL agent's performance. Information about the influence of these experiences is valuable for various purposes, such as identifying experiences that negatively influence poorly performing RL agents. One method for estimating the influence of experiences is the leave-one-out (LOO) method. However, this method is usually computationally prohibitive. In this paper, we present Policy Iteration with Turn-over Dropout (PIToD), which efficiently estimates the influence of experiences. We evaluate how accurately PIToD estimates the influence of experiences and its efficiency compared to LOO. We then apply PIToD to amend poorly performing RL agents, i.e., we use PIToD to estimate negatively influential experiences for the RL agents and to delete the influence of these experiences. We show that RL agents' performance is significantly improved via amendments with PIToD., Comment: Source code: https://github.com/TakuyaHiraoka/Which-Experiences-Are-Influential-for-RL-Agents

Library-based methods are known to be very effective for fast motion planning by adapting an experience retrieved from a precomputed library. This article presents CoverLib, a principled approach for constructing and utilizing such a library. CoverLib iteratively adds an experience-classifier-pair to the library, where each classifier corresponds to an adaptable region of the experience within the problem space. This iterative process is an active procedure, as it selects the next experience based on its ability to effectively cover the uncovered region. During the query phase, these classifiers are utilized to select an experience that is expected to be adaptable for a given problem. Experimental results demonstrate that CoverLib effectively mitigates the trade-off between plannability and speed observed in global (e.g. sampling-based) and local (e.g. optimization-based) methods. As a result, it achieves both fast planning and high success rates over the problem domain. Moreover, due to its adaptation-algorithm-agnostic nature, CoverLib seamlessly integrates with various adaptation methods, including nonlinear programming-based and sampling-based algorithms.

The stability of persistent homology has led to wide applications of the persistence diagram as a trusted topological descriptor in the presence of noise. However, with the increasing demand for high-dimension and low-sample-size data processing in modern science, it is questionable whether persistence diagrams retain their reliability in the presence of high-dimensional noise. This work aims to study the reliability of persistence diagrams in the high-dimension low-sample-size data setting. By analyzing the asymptotic behavior of persistence diagrams for high-dimensional random data, we show that persistence diagrams are no longer reliable descriptors of low-sample-size data under high-dimensional noise perturbations. We refer to this loss of reliability of persistence diagrams in such data settings as the curse of dimensionality on persistence diagrams. Next, we investigate the possibility of using normalized principal component analysis as a method for reducing the dimensionality of the high-dimensional observed data to resolve the curse of dimensionality. We show that this method can mitigate the curse of dimensionality on persistence diagrams. Our results shed some new light on the challenges of processing high-dimension low-sample-size data by persistence diagrams and provide a starting point for future research in this area.

We explore threshold vocabulary trimming in Byte-Pair Encoding subword tokenization, a postprocessing step that replaces rare subwords with their component subwords. The technique is available in popular tokenization libraries but has not been subjected to rigorous scientific scrutiny. While the removal of rare subwords is suggested as best practice in machine translation implementations, both as a means to reduce model size and for improving model performance through robustness, our experiments indicate that, across a large space of hyperparameter settings, vocabulary trimming fails to improve performance, and is even prone to incurring heavy degradation., Comment: 15 pages

Large language models (LLMs) have increased interest in vision language models (VLMs), which process image-text pairs as input. Studies investigating the visual understanding ability of VLMs have been proposed, but such studies are still preliminary because existing datasets do not permit a comprehensive evaluation of the fine-grained visual linguistic abilities of VLMs across multiple languages. To further explore the strengths of VLMs, such as GPT-4V \cite{openai2023GPT4}, we developed new datasets for the systematic and qualitative analysis of VLMs. Our contribution is four-fold: 1) we introduced nine vision-and-language (VL) tasks (including object recognition, image-text matching, and more) and constructed multilingual visual-text datasets in four languages: English, Japanese, Swahili, and Urdu through utilizing templates containing \textit{questions} and prompting GPT4-V to generate the \textit{answers} and the \textit{rationales}, 2) introduced a new VL task named \textit{unrelatedness}, 3) introduced rationales to enable human understanding of the VLM reasoning process, and 4) employed human evaluation to measure the suitability of proposed datasets for VL tasks. We show that VLMs can be fine-tuned on our datasets. Our work is the first to conduct such analyses in Swahili and Urdu. Also, it introduces \textit{rationales} in VL analysis, which played a vital role in the evaluation.

Do pretrained language models have knowledge regarding the surface information of tokens? We examined the surface information stored in word or subword embeddings acquired by pretrained language models from the perspectives of token length, substrings, and token constitution. Additionally, we evaluated the ability of models to generate knowledge regarding token surfaces. We focused on 12 pretrained language models that were mainly trained on English and Japanese corpora. Experimental results demonstrate that pretrained language models have knowledge regarding token length and substrings but not token constitution. Additionally, the results imply that there is a bottleneck on the decoder side in terms of effectively utilizing acquired knowledge.

Transferring human motion skills to humanoid robots remains a significant challenge. In this study, we introduce a Wasserstein adversarial imitation learning system, allowing humanoid robots to replicate natural whole-body locomotion patterns and execute seamless transitions by mimicking human motions. First, we present a unified primitive-skeleton motion retargeting to mitigate morphological differences between arbitrary human demonstrators and humanoid robots. An adversarial critic component is integrated with Reinforcement Learning (RL) to guide the control policy to produce behaviors aligned with the data distribution of mixed reference motions. Additionally, we employ a specific Integral Probabilistic Metric (IPM), namely the Wasserstein-1 distance with a novel soft boundary constraint to stabilize the training process and prevent mode collapse. Our system is evaluated on a full-sized humanoid JAXON in the simulator. The resulting control policy demonstrates a wide range of locomotion patterns, including standing, push-recovery, squat walking, human-like straight-leg walking, and dynamic running. Notably, even in the absence of transition motions in the demonstration dataset, robots showcase an emerging ability to transit naturally between distinct locomotion patterns as desired speed changes.

We performed time-resolved pump--probe measurements using rare-earth iron garnet \ce{Gd3/2Yb1/2BiFe5O12} as a two-sublattice ferrimagnet. We measured the initial phases of the magnetic resonance modes below and above the magnetization compensation temperature to clarify the sublattice selectivity of the inverse Faraday effect in ferrimagnets. A comparison of the time evolution of magnetization estimated using the equations of motion revealed that the inverse Faraday effect occurring in ferrimagnetic materials has sublattice selectivity. This is in striking contrast to antiferromagnets, in which the inverse Faraday effect acts on each sublattice identically. The initial phase analysis can be applied to other ferrimagnets with compensation temperatures., Comment: 4 pages, 5 figures

Reinforcement learning (RL) methods with a high replay ratio (RR) and regularization have gained interest due to their superior sample efficiency. However, these methods have mainly been developed for dense-reward tasks. In this paper, we aim to extend these RL methods to sparse-reward goal-conditioned tasks. We use Randomized Ensemble Double Q-learning (REDQ) (Chen et al., 2021), an RL method with a high RR and regularization. To apply REDQ to sparse-reward goal-conditioned tasks, we make the following modifications to it: (i) using hindsight experience replay and (ii) bounding target Q-values. We evaluate REDQ with these modifications on 12 sparse-reward goal-conditioned tasks of Robotics (Plappert et al., 2018), and show that it achieves about $2 \times$ better sample efficiency than previous state-of-the-art (SoTA) RL methods. Furthermore, we reconsider the necessity of specific components of REDQ and simplify it by removing unnecessary ones. The simplified REDQ with our modifications achieves $\sim 8 \times$ better sample efficiency than the SoTA methods in 4 Fetch tasks of Robotics., Comment: Source code: https://github.com/TakuyaHiraoka/Efficient-SRGC-RL-with-a-High-RR-and-Regularization Demo video: https://drive.google.com/file/d/1UHd7JVPCwFLNFhy1QcycQfwU_nll_yII/view?usp=drive_link

A fundamental challenge in multiparameter persistent homology is the absence of a complete and discrete invariant. To address this issue, we propose an enhanced framework that realizes a holistic understanding of a fully commutative quiver's representation via synthesizing interpretations obtained from intervals. Additionally, it provides a mechanism to tune the balance between approximation resolution and computational complexity. This framework is evaluated on commutative ladders of both finite-type and infinite-type. For the former, we discover an efficient method for the indecomposable decomposition leveraging solely one-parameter persistent homology. For the latter, we introduce a new invariant that reveals persistence in the second parameter by connecting two standard persistence diagrams using interval approximations. We subsequently present several models for constructing commutative ladder filtrations, offering fresh insights into random filtrations and demonstrating our toolkit's effectiveness in analyzing the topology of materials.

Imitation learning has been actively studied in recent years. In particular, skill acquisition by a robot with a fixed body, whose root link position and posture and camera angle of view do not change, has been realized in many cases. On the other hand, imitation of the behavior of robots with floating links, such as humanoid robots, is still a difficult task. In this study, we develop an imitation learning system using a biped robot with a floating link. There are two main problems in developing such a system. The first is a teleoperation device for humanoids, and the second is a control system that can withstand heavy workloads and long-term data collection. For the first point, we use the whole body control device TABLIS. It can control not only the arms but also the legs and can perform bilateral control with the robot. By connecting this TABLIS with the high-power humanoid robot JAXON, we construct a control system for imitation learning. For the second point, we will build a system that can collect long-term data based on posture optimization, and can simultaneously move the robot's limbs. We combine high-cycle posture generation with posture optimization methods, including whole-body joint torque minimization and contact force optimization. We designed an integrated system with the above two features to achieve various tasks through imitation learning. Finally, we demonstrate the effectiveness of this system by experiments of manipulating flexible fabrics such that not only the hands but also the head and waist move simultaneously, manipulating objects using legs characteristic of humanoids, and lifting heavy objects that require large forces., Comment: accepted at IROS2023

Node2vec is a graph embedding method that learns a vector representation for each node of a weighted graph while seeking to preserve relative proximity and global structure. Numerical experiments suggest Node2vec struggles to recreate the topology of the input graph. To resolve this we introduce a topological loss term to be added to the training loss of Node2vec which tries to align the persistence diagram (PD) of the resulting embedding as closely as possible to that of the input graph. Following results in computational optimal transport, we carefully adapt entropic regularization to PD metrics, allowing us to measure the discrepancy between PDs in a differentiable way. Our modified loss function can then be minimized through gradient descent to reconstruct both the geometry and the topology of the input graph. We showcase the benefits of this approach using demonstrative synthetic examples., Comment: For associated repository, see https://github.com/killianfmeehan/topological_node2vec

When a fraction of a population becomes immune to an infectious disease, the population-wide infection risk decreases nonlinearly due to collective protection known as herd immunity. Studies based on mean-field models suggest that natural infection in a heterogeneous population may induce herd immunity more efficiently than homogeneous immunization. Here, we use network epidemic models to show that the opposite can also be the case. We identify two competing mechanisms driving disease-induced herd immunity in networks: the high density of immunity among socially active individuals enhances the herd immunity effect, while the topological localization of immune individuals weakens it. The effect of localization is stronger in networks embedded in low-dimensional space, which can make disease-induced immunity less effective than random immunization. Our results highlight the role of networks in shaping herd immunity and call for careful examination of model predictions that inform public health policies., Comment: Main text: 11 pages, 4 figures. Supplementary Materials: 8 pages, 2 figures

We explored quinary body-centered cubic (bcc) high-entropy alloy (HEA) superconductors with valence electron concentrations (VECs) ranging from 4.6 to 5.0, a domain that has received limited attention in prior research. Our search has led to the discovery of new bcc Ti-Hf-Nb-Ta-Re superconducting alloys, which exhibit an interesting phenomenon of phase segregation into two bcc phases with slightly different chemical compositions, as the VEC increases. The enthalpy of the formation of each binary compound explains the phase segregation. All the alloys investigated were categorized as type-II superconductors, with superconducting critical temperatures ($T_\mathrm{c}$) ranging from 3.25 K to 4.38 K. We measured the Vickers microhardness, which positively correlated with the Debye temperature, and compared it with the hardness values of other bcc HEA superconductors. Our results indicate that $T_\mathrm{c}$ systematically decreases with an increase in hardness beyond a threshold of approximately 350 HV. Additionally, we plotted $T_\mathrm{c}$ vs. VEC for representative quinary bcc HEAs. The plot revealed the asymmetric VEC dependence. The correlation between the hardness and $T_\mathrm{c}$, as well as the asymmetric dependence of $T_\mathrm{c}$ on VEC can be attributed to the simultaneous effects of the electronic density of states at the Fermi level and electron-phonon coupling under the uncertainty principle, especially in the higher VEC region., Comment: to appear in Journal of Alloys and Metallurgical Systems

Xenon dual-phase time projections chambers (TPCs) have proven to be a successful technology in studying physical phenomena that require low-background conditions. With 40t of liquid xenon (LXe) in the TPC baseline design, DARWIN will have a high sensitivity for the detection of particle dark matter, neutrinoless double beta decay ($0\nu\beta\beta$), and axion-like particles (ALPs). Although cosmic muons are a source of background that cannot be entirely eliminated, they may be greatly diminished by placing the detector deep underground. In this study, we used Monte Carlo simulations to model the cosmogenic background expected for the DARWIN observatory at four underground laboratories: Laboratori Nazionali del Gran Sasso (LNGS), Sanford Underground Research Facility (SURF), Laboratoire Souterrain de Modane (LSM) and SNOLAB. We determine the production rates of unstable xenon isotopes and tritium due to muon-included neutron fluxes and muon-induced spallation. These are expected to represent the dominant contributions to cosmogenic backgrounds and thus the most relevant for site selection.

Level 3 automated driving systems (ADS) have attracted significant attention and are being commercialized. A Level 3 ADS prompts the driver to take control by requesting to intervene (RtI) when its operational design domain (ODD) or system limitations are exceeded. However, complex traffic situations may lead drivers to perceive multiple potential triggers of RtI simultaneously, causing hesitation or confusion during take-over. Therefore, drivers must clearly understand the ADS's system limitations to understand the triggers of RtI and ensure safe take-over. In this study, we propose a voice-based HMI for providing RtI trigger cues to help drivers understand ADS's system limitations. The results of a between-group experiment using a driving simulator showed that incorporating effective trigger cues into the RtI enabled drivers to comprehend the ADS's system limitations better and reduce collisions. It also improved the subjective evaluations of drivers, such as the comprehensibility of system limitations, hesitation in response to RtI, and acceptance of ADS behaviors when encountering RtI while using the ADS. Therefore, enhanced comprehension resulting from trigger cues is essential for promoting a safer and better user experience using ADS during RtI.

We have synthesized new Tsai-type Yb-based intermediate-valence approximant crystals (ACs) with different degree of approximation to quasicrystal, Zn--Au--Yb 1/1 and 2/1 AC, and studied the external pressure effect on their Yb mean-valence $\nu$. Whereas 1/1 AC distinctly exhibits a first-order-like jump in $\nu$ at a transition pressure $P_{\rm v}$, 2/1 AC only shows an indistinct anomaly at $P_{\rm v}$. We have also studied the pressure dependence of the $\nu$ of Au--Al--Yb 1/1 AC, which is a prototypal AC exhibiting pressure-induced quantum criticality. It shows a continuous valence anomaly at a critical pressure $P_{\rm c}$ where the magnetic susceptibility diverges toward zero temperature, in contrast to the valence jump in the Zn--Au--Yb 1/1 AC. These results are discussed based on a theoretical model of quantum critical valence fluctuation.

Recently, methods for learning diverse skills to generate various behaviors without external rewards have been actively studied as a form of unsupervised reinforcement learning. However, most of the existing methods learn a finite number of discrete skills, and thus the variety of behaviors that can be exhibited with the learned skills is limited. In this paper, we propose a novel method for learning potentially an infinite number of different skills, which is named discovery of continuous skills on a sphere (DISCS). In DISCS, skills are learned by maximizing mutual information between skills and states, and each skill corresponds to a continuous value on a sphere. Because the representations of skills in DISCS are continuous, infinitely diverse skills could be learned. We examine existing methods and DISCS in the MuJoCo Ant robot control environments and show that DISCS can learn much more diverse skills than the other methods., Comment: 14 pages, 12 figures

Is preferred tokenization for humans also preferred for machine-learning (ML) models? This study examines the relations between preferred tokenization for humans (appropriateness and readability) and one for ML models (performance on an NLP task). The question texts of the Japanese commonsense question-answering dataset are tokenized with six different tokenizers, and the performances of human annotators and ML models were compared. Furthermore, we analyze relations among performance of answers by human and ML model, the appropriateness of tokenization for human, and response time to questions by human. This study provides a quantitative investigation result that shows that preferred tokenizations for humans and ML models are not necessarily always the same. The result also implies that existing methods using language models for tokenization could be a good compromise both for human and ML models.

This paper proposes a method to optimize tokenization for the performance improvement of already trained downstream models. Our method generates tokenization results attaining lower loss values of a given downstream model on the training data for restricting vocabularies and trains a tokenizer reproducing the tokenization results. Therefore, our method can be applied to variety of tokenization methods, while existing work cannot due to the simultaneous learning of the tokenizer and the downstream model. This paper proposes an example of the BiLSTM-based tokenizer with vocabulary restriction, which can capture wider contextual information for the tokenization process than non-neural-based tokenization methods used in existing work. Experimental results on text classification in Japanese, Chinese, and English text classification tasks show that the proposed method improves performance compared to the existing methods for tokenization optimization.