Yutaka Matsuo

IJCAI Conference 2025 Conference Paper

A Comprehensive Survey on Physical Risk Control in the Era of Foundation Model-enabled Robotics

• Takeshi Kojima
• Yaonan Zhu
• Yusuke Iwasawa
• Toshinori Kitamura
• Gang Yan
• Shu Morikuni
• Ryosuke Takanami
• Alfredo Solano

Recent Foundation Model-enabled robotics (FMRs) display greatly improved general-purpose skills, enabling more adaptable automation than conventional robotics. Their ability to handle diverse tasks thus creates new opportunities to replace human labor. However, unlike general foundation models, FMRs interact with the physical world, where their actions directly affect the safety of humans and surrounding objects, requiring careful deployment and control. Based on this proposition, our survey comprehensively summarizes robot control approaches to mitigate physical risks by covering all the lifespan of FMRs ranging from pre-deployment to post-accident stage. Specifically, we broadly divide the timeline into the following three phases: (1) pre-deployment phase, (2) pre-incident phase, and (3) post-incident phase. Throughout this survey, we find that there is much room to study (i) pre-incident risk mitigation strategies, (ii) research that assumes physical interaction with humans, and (iii) essential issues of foundation models themselves. We hope that this survey will be a milestone in providing a high-resolution analysis of the physical risks of FMRs and their control, contributing to the realization of a good human-robot relationship.

ICML Conference 2025 Conference Paper

Beyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit Emergence

• Gouki Minegishi
• Hiroki Furuta
• Shohei Taniguchi
• Yusuke Iwasawa
• Yutaka Matsuo

Transformer-based language models exhibit In-Context Learning (ICL), where predictions are made adaptively based on context. While prior work links induction heads to ICL through a sudden jump in accuracy, this can only account for ICL when the answer is included within the context. However, an important property of practical ICL in large language models is the ability to meta-learn how to solve tasks from context, rather than just copying answers from context; how such an ability is obtained during training is largely unexplored. In this paper, we experimentally clarify how such meta-learning ability is acquired by analyzing the dynamics of the model’s circuit during training. Specifically, we extend the copy task from previous research into an In-Context Meta Learning setting, where models must infer a task from examples to answer queries. Interestingly, in this setting, we find that there are multiple phases in the process of acquiring such abilities, and that a unique circuit emerges in each phase, contrasting with the single-phases change in induction heads. The emergence of such circuits can be related to several phenomena known in large language models, and our analysis lead to a deeper understanding of the source of the transformer’s ICL ability.

TMLR Journal 2025 Journal Article

Bridging Lottery Ticket and Grokking: Understanding Grokking from Inner Structure of Networks

• Gouki Minegishi
• Yusuke Iwasawa
• Yutaka Matsuo

Grokking is an intriguing phenomenon of delayed generalization, where neural networks initially memorize training data with perfect accuracy but exhibit poor generalization, subsequently transitioning to a generalizing solution with continued training. While factors such as weight norms and sparsity have been proposed to explain this delayed generalization, the influence of network structure remains underexplored. In this work, we link the grokking phenomenon to the lottery ticket hypothesis to investigate the impact of internal network structures. We demonstrate that utilizing lottery tickets obtained during the generalizing phase (termed grokked tickets) significantly reduces delayed generalization across various tasks, including multiple modular arithmetic operations, polynomial regression, sparse parity, and MNIST classification. Through controlled experiments, we show that the mitigation of delayed generalization is not due solely to reduced weight norms or increased sparsity, but rather to the discovery of good subnetworks. Furthermore, we find that grokked tickets exhibit periodic weight patterns and undergo rapid structural changes that coincide with improvements in generalization. Additionally, pruning techniques like the edge-popup algorithm can identify these effective structures without modifying the weights, thereby transforming memorizing networks into generalizing ones. These results underscore the novel insight that structural exploration plays a pivotal role in understanding grokking.

TMLR Journal 2025 Journal Article

Efficient Object-Centric Representation Learning using Masked Generative Modeling

• Akihiro Nakano
• Masahiro Suzuki
• Yutaka Matsuo

Learning object-centric representations from visual inputs in an unsupervised manner has drawn focus to solve more complex tasks, such as reasoning and reinforcement learning. However, current state-of-the-art methods, relying on autoregressive transformers or diffusion models to generate scenes from object-centric representations, suffer from computational inefficiency due to their sequential or iterative nature. This computational bottleneck limits their practical application and hinders scaling to more complex downstream tasks. To overcome this, we propose MOGENT, an efficient object-centric learning framework based on masked generative modeling. MOGENT conditions a masked bidirectional transformer on learned object slots and employs a parallel iterative decoding scheme to generate scenes, enabling efficient compositional generation. Experiments show that MOGENT significantly improves computational efficiency, accelerating the generation process by up to 67x and 17x compared to autoregressive models and diffusion-based models, respectively. Importantly, the efficiency is attained while maintaining strong or competitive performance on object segmentation and compositional generation tasks.

NeurIPS Conference 2025 Conference Paper

Inference-Time Text-to-Video Alignment with Diffusion Latent Beam Search

• Yuta Oshima
• Masahiro Suzuki
• Yutaka Matsuo
• Hiroki Furuta

The remarkable progress in text-to-video diffusion models enables the generation of photorealistic videos, although the content of these generated videos often includes unnatural movement or deformation, reverse playback, and motionless scenes. Recently, an alignment problem has attracted huge attention, where we steer the output of diffusion models based on some measure of the content's goodness. Because there is a large room for improvement of perceptual quality along the frame direction, we should address which metrics we should optimize and how we can optimize them in the video generation. In this paper, we propose diffusion latent beam search with lookahead estimator, which can select a better diffusion latent to maximize a given alignment reward at inference time. We then point out that improving perceptual video quality with respect to alignment to prompts requires reward calibration by weighting existing metrics. This is because when humans or vision language models evaluate outputs, many previous metrics to quantify the naturalness of video do not always correlate with the evaluation. We demonstrate that our method improves the perceptual quality evaluated on the calibrated reward, VLMs, and human assessment, without model parameter update, and outputs the best generation compared to greedy search and best-of-N sampling under much more efficient computational cost. The experiments highlight that our method is beneficial to many capable generative models, and provide a practical guideline: we should prioritize the inference-time compute allocation into enabling the lookahead estimator and increasing the search budget, rather than expanding the denoising steps.

ICLR Conference 2025 Conference Paper

Near-Optimal Policy Identification in Robust Constrained Markov Decision Processes via Epigraph Form

• Toshinori Kitamura
• Tadashi Kozuno
• Wataru Kumagai
• Kenta Hoshino
• Yohei Hosoe
• Kazumi Kasaura
• Masashi Hamaya
• Paavo Parmas

Designing a safe policy for uncertain environments is crucial in real-world control systems. However, this challenge remains inadequately addressed within the Markov decision process (MDP) framework. This paper presents the first algorithm guaranteed to identify a near-optimal policy in a robust constrained MDP (RCMDP), where an optimal policy minimizes cumulative cost while satisfying constraints in the worst-case scenario across a set of environments. We first prove that the conventional policy gradient approach to the Lagrangian max-min formulation can become trapped in suboptimal solutions. This occurs when its inner minimization encounters a sum of conflicting gradients from the objective and constraint functions. To address this, we leverage the epigraph form of the RCMDP problem, which resolves the conflict by selecting a single gradient from either the objective or the constraints. Building on the epigraph form, we propose a bisection search algorithm with a policy gradient subroutine and prove that it identifies an $\varepsilon$-optimal policy in an RCMDP with $\widetilde{\mathcal{O}}(\varepsilon^{-4})$ robust policy evaluations.

NeurIPS Conference 2025 Conference Paper

Provably Efficient RL under Episode-Wise Safety in Constrained MDPs with Linear Function Approximation

• Toshinori Kitamura
• Arnob Ghosh
• Tadashi Kozuno
• Wataru Kumagai
• Kazumi Kasaura
• Kenta Hoshino
• Yohei Hosoe
• Yutaka Matsuo

We study the reinforcement learning (RL) problem in a constrained Markov decision process (CMDP), where an agent explores the environment to maximize the expected cumulative reward while satisfying a single constraint on the expected total utility value in every episode. While this problem is well understood in the tabular setting, theoretical results for function approximation remain scarce. This paper closes the gap by proposing an RL algorithm for linear CMDPs that achieves $\widetilde{\mathcal{O}}(\sqrt{K})$ regret with an episode-wise zero-violation guarantee. Furthermore, our method is computationally efficient, scaling polynomially with problem-dependent parameters while remaining independent of the state space size. Our results significantly improve upon recent linear CMDP algorithms, which either violate the constraint or incur exponential computational costs.

ICLR Conference 2025 Conference Paper

Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words

• Gouki Minegishi
• Hiroki Furuta
• Yusuke Iwasawa
• Yutaka Matsuo

Sparse autoencoders (SAEs) have gained a lot of attention as a promising tool to improve the interpretability of large language models (LLMs) by mapping the complex superposition of *polysemantic* neurons into *monosemantic* features and composing a sparse dictionary of words. However, traditional performance metrics like Mean Squared Error and $\mathrm{L}_{0}$ sparsity ignore the evaluation of the semantic representational power of SAEs - whether they can acquire interpretable monosemantic features while preserving the semantic relationship of words.For instance, it is not obvious whether a learned sparse feature could distinguish different meanings in one word. In this paper, we propose a suite of evaluations for SAEs to analyze the quality of monosemantic features by focusing on polysemous words. Our findings reveal that SAEs developed to improve the MSE-$\mathrm{L}_0$ Pareto frontier may confuse interpretability, which does not necessarily enhance the extraction of monosemantic features. The analysis of SAEs with polysemous words can also figure out the internal mechanism of LLMs; deeper layers and the Attention module contribute to distinguishing polysemy in a word. Our semantics-focused evaluation offers new insights into the polysemy and the existing SAE objective and contributes to the development of more practical SAEs.

TMLR Journal 2025 Journal Article

The Geometry of Phase Transitions in Diffusion Models: Tubular Neighbourhoods and Singularities

• Manato Yaguchi
• Kotaro Sakamoto
• Ryosuke Sakamoto
• Masato Tanabe
• Masatomo Akagawa
• Yusuke Hayashi
• Masahiro Suzuki
• Yutaka Matsuo

Diffusion models undergo phase transitions during the generative process where data features suddenly emerge in the final stages. The current study aims to elucidate this critical phenomenon from the geometrical perspective. We employ the concept of ``injectivity radius'', a quantity that characterises the structure of the data manifold. Through theoretical and empirical evidence, we demonstrate that phase transitions in the generative process of diffusion models are closely related to the injectivity radius. Our findings offer a novel perspective on phase transitions in diffusion models, with potential implications for improving performance and sampling efficiency.

NeurIPS Conference 2025 Conference Paper

Topology of Reasoning: Understanding Large Reasoning Models through Reasoning Graph Properties

• Gouki Minegishi
• Hiroki Furuta
• Takeshi Kojima
• Yusuke Iwasawa
• Yutaka Matsuo

Recent large-scale reasoning models have achieved state-of-the-art performance on challenging mathematical benchmarks, yet the internal mechanisms underlying their success remain poorly understood. In this work, we introduce the notion of a reasoning graph, extracted by clustering hidden‐state representations at each reasoning step, and systematically analyze three key graph-theoretic properties: cyclicity, diameter, and small-world index, across multiple tasks (GSM8K, MATH500, AIME~2024). Our findings reveal that distilled reasoning models (e. g. , DeepSeek-R1-Distill-Qwen-32B) exhibit significantly more recurrent cycles (about 5 per sample), substantially larger graph diameters, and pronounced small-world characteristics (about 6x) compared to their base counterparts. Notably, these structural advantages grow with task difficulty and model capacity, with cycle detection peaking at the 14B scale and exploration diameter maximized in the 32B variant, correlating positively with accuracy. Furthermore, we show that supervised fine-tuning on an improved dataset systematically expands reasoning graph diameters in tandem with performance gains, offering concrete guidelines for dataset design aimed at boosting reasoning capabilities. By bridging theoretical insights into reasoning graph structures with practical recommendations for data construction, our work advances both the interpretability and the efficacy of large reasoning models.

ICLR Conference 2024 Conference Paper

A Real-World WebAgent with Planning, Long Context Understanding, and Program Synthesis

• Izzeddin Gur
• Hiroki Furuta
• Austin V. Huang
• Mustafa Safdari
• Yutaka Matsuo
• Douglas Eck
• Aleksandra Faust

Pre-trained large language models (LLMs) have recently achieved better generalization and sample efficiency in autonomous web automation. However, the performance on real-world websites has still suffered from (1) open domainness, (2) limited context length, and (3) lack of inductive bias on HTML. We introduce WebAgent, an LLM-driven agent that learns from self-experience to complete tasks on real websites following natural language instructions. WebAgent plans ahead by decomposing instructions into canonical sub-instructions, summarizes long HTML documents into task-relevant snippets, and acts on websites via Python programs generated from those. We design WebAgent with Flan-U-PaLM, for grounded code generation, and HTML-T5, new pre-trained LLMs for long HTML documents using local and global attention mechanisms and a mixture of long-span denoising objectives, for planning and summarization. We empirically demonstrate that our modular recipe improves the success on real websites by over 50%, and that HTML-T5 is the best model to solve various HTML understanding tasks; achieving 18.7% higher success rate than the prior method on MiniWoB web automation benchmark, and SoTA performance on Mind2Web, an offline task planning evaluation.

NeurIPS Conference 2024 Conference Paper

ADOPT: Modified Adam Can Converge with Any $\beta_2$ with the Optimal Rate

• Shohei Taniguchi
• Keno Harada
• Gouki Minegishi
• Yuta Oshima
• Seong Cheol Jeong
• Go Nagahara
• Tomoshi Iiyama
• Masahiro Suzuki

Adam is one of the most popular optimization algorithms in deep learning. However, it is known that Adam does not converge in theory unless choosing a hyperparameter, i. e. , $\beta_2$, in a problem-dependent manner. There have been many attempts to fix the non-convergence (e. g. , AMSGrad), but they require an impractical assumption that the gradient noise is uniformly bounded. In this paper, we propose a new adaptive gradient method named ADOPT, which achieves the optimal convergence rate of $\mathcal{O} ( 1 / \sqrt{T} )$ with any choice of $\beta_2$ without depending on the bounded noise assumption. ADOPT addresses the non-convergence issue of Adam by removing the current gradient from the second moment estimate and changing the order of the momentum update and the normalization by the second moment estimate. We also conduct intensive numerical experiments, and verify that our ADOPT achieves superior results compared to Adam and its variants across a wide range of tasks, including image classification, generative modeling, natural language processing, and deep reinforcement learning. The implementation is available at https: //github. com/iShohei220/adopt.

TMLR Journal 2024 Journal Article

Exposing Limitations of Language Model Agents in Sequential-Task Compositions on the Web

• Hiroki Furuta
• Yutaka Matsuo
• Aleksandra Faust
• Izzeddin Gur

Language model agents (LMA) recently emerged as a promising paradigm on muti-step decision making tasks, often outperforming humans and other reinforcement learning agents. Despite the promise, their performance on real-world applications that often involve combinations of tasks is still underexplored. In this work, we introduce a new benchmark, called CompWoB -- 50 new compositional web automation tasks reflecting more realistic assumptions. We show that while existing prompted LMAs (gpt-3.5-turbo or gpt-4) achieve 94.0% average success rate on base tasks, their performance degrades to 24.9% success rate on compositional tasks. On the other hand, transferred LMAs (finetuned only on base tasks) show less generalization gap, dropping from 85.4% to 54.8%. By balancing data distribution across tasks, we train a new model, HTML-T5++, that surpasses human-level performance (95.2%) on MiniWoB, and achieves the best zero-shot performance on CompWoB (61.5%). While these highlight the promise of small-scale finetuned and transferred models for task compositionality, their performance further degrades under different instruction compositions changing combinational order. In contrast to the recent remarkable success of LMA, our benchmark and detailed analysis emphasize the necessity of building LMAs that are robust and generalizable to task compositionality for real-world deployment.

ICRA Conference 2024 Conference Paper

GenDOM: Generalizable One-shot Deformable Object Manipulation with Parameter-Aware Policy

• So Kuroki
• Jiaxian Guo
• Tatsuya Matsushima
• Takuya Okubo
• Masato Kobayashi 0001
• Yuya Ikeda
• Ryosuke Takanami
• Paul Yoo

Due to the inherent uncertainty in their deformability during motion, previous methods in deformable object manipulation, such as rope and cloth, often required hundreds of real-world demonstrations to train a manipulation policy for each object, which hinders their applications in our ever-changing world. To address this issue, we introduce GenDOM, a framework that allows the manipulation policy to handle different deformable objects with only a single real-world demonstration. To achieve this, we augment the policy by conditioning it on deformable object parameters and training it with a diverse range of simulated deformable objects so that the policy can adjust actions based on different object parameters. At the time of inference, given a new object, GenDOM can estimate the deformable object parameters with only a single real-world demonstration by minimizing the disparity between the grid density of point clouds of real-world demonstrations and simulations in a differentiable physics simulator. Empirical validations on both simulated and real-world object manipulation setups clearly show that our method can manipulate different objects with a single demonstration and significantly outperforms the baseline in both environments (a 62% improvement for in-domain ropes and a 15% improvement for out-of-distribution ropes in simulation, as well as a 26% improvement for ropes and a 50% improvement for cloths in the real world), demonstrating the effectiveness of our approach in one-shot deformable object manipulation. https://sites.google.com/view/gendom/home.

NeurIPS Conference 2024 Conference Paper

Geometric-Averaged Preference Optimization for Soft Preference Labels

• Hiroki Furuta
• Kuang-Huei Lee
• Shixiang S. Gu
• Yutaka Matsuo
• Aleksandra Faust
• Heiga Zen
• Izzeddin Gur

Many algorithms for aligning LLMs with human preferences assume that human preferences are binary and deterministic. However, human preferences can vary across individuals, and therefore should be represented distributionally. In this work, we introduce the distributional soft preference labels and improve Direct Preference Optimization (DPO) with a weighted geometric average of the LLM output likelihood in the loss function. This approach adjusts the scale of learning loss based on the soft labels such that the loss would approach zero when the responses are closer to equally preferred. This simple modification can be easily applied to any DPO-based methods and mitigate over-optimization and objective mismatch, which prior works suffer from. Our experiments simulate the soft preference labels with AI feedback from LLMs and demonstrate that geometric averaging consistently improves performance on standard benchmarks for alignment research. In particular, we observe more preferable responses than binary labels and significant improvements where modestly-confident labels are in the majority.

ICLR Conference 2024 Conference Paper

Multimodal Web Navigation with Instruction-Finetuned Foundation Models

• Hiroki Furuta
• Kuang-Huei Lee
• Ofir Nachum
• Yutaka Matsuo
• Aleksandra Faust
• Shixiang Gu
• Izzeddin Gur

The progress of autonomous web navigation has been hindered by the dependence on billions of exploratory interactions via online reinforcement learning, and domain-specific model designs that make it difficult to leverage generalization from rich out-of-domain data. In this work, we study data-driven offline training for web agents with vision-language foundation models. We propose an instruction-following multimodal agent, WebGUM, that observes both webpage screenshots and HTML pages and outputs web navigation actions, such as click and type. WebGUM is trained by jointly finetuning an instruction-finetuned language model and a vision encoder with temporal and local perception on a large corpus of demonstrations. We empirically demonstrate this recipe improves the agent's ability of grounded multimodal perception, HTML comprehension, and multi-step reasoning, outperforming prior works by a significant margin. On the MiniWoB, we improve over the previous best offline methods by more than 45.8%, even outperforming online-finetuned SoTA, humans, and GPT-4-based agent. On the WebShop benchmark, our 3-billion-parameter model achieves superior performance to the existing SoTA, PaLM-540B. Furthermore, WebGUM exhibits strong positive transfer to the real-world planning tasks on the Mind2Web. We also collect 347K high-quality demonstrations using our trained models, 38 times larger than prior work, and make them available to promote future research in this direction.

ICRA Conference 2024 Conference Paper

Open X-Embodiment: Robotic Learning Datasets and RT-X Models: Open X-Embodiment Collaboration

• Abby O'Neill
• Abdul Rehman
• Abhiram Maddukuri
• Abhishek Gupta 0004
• Abhishek Padalkar
• Abraham Lee
• Acorn Pooley
• Agrim Gupta

Large, high-capacity models trained on diverse datasets have shown remarkable successes on efficiently tackling downstream applications. In domains from NLP to Computer Vision, this has led to a consolidation of pretrained models, with general pretrained backbones serving as a starting point for many applications. Can such a consolidation happen in robotics? Conventionally, robotic learning methods train a separate model for every application, every robot, and even every environment. Can we instead train "generalist" X-robot policy that can be adapted efficiently to new robots, tasks, and environments? In this paper, we provide datasets in standardized data formats and models to make it possible to explore this possibility in the context of robotic manipulation, alongside experimental results that provide an example of effective X-robot policies. We assemble a dataset from 22 different robots collected through a collaboration between 21 institutions, demonstrating 527 skills (160266 tasks). We show that a high-capacity model trained on this data, which we call RT-X, exhibits positive transfer and improves the capabilities of multiple robots by leveraging experience from other platforms. The project website is robotics-transformer-x. github.io.

ICRA Conference 2024 Conference Paper

Self-Recovery Prompting: Promptable General Purpose Service Robot System with Foundation Models and Self-Recovery

• Mimo Shirasaka
• Tatsuya Matsushima
• Soshi Tsunashima
• Yuya Ikeda
• Aoi Horo
• So Ikoma
• Chikaha Tsuji
• Hikaru Wada

A general-purpose service robot (GPSR), which can execute diverse tasks in various environments, requires a system with high generalizability and adaptability to tasks and environments. In this paper, we first developed a top-level GPSR system for worldwide competition (RoboCup@Home2023) based on multiple foundation models. This system is both generalizable to variations and adaptive by prompting each model. Then, by analyzing the performance of the developed system, we found three types of failure in more realistic GPSR application settings: insufficient information, incorrect plan generation, and plan execution failure. We then propose the self-recovery prompting pipeline, which explores the necessary information and modifies its prompts to recover from failure. We experimentally confirm that the system with the self-recovery mechanism can accomplish tasks by resolving various failure cases. https://sites.google.com/view/srgpsr

TMLR Journal 2024 Journal Article

Towards Empirical Interpretation of Internal Circuits and Properties in Grokked Transformers on Modular Polynomials

• Hiroki Furuta
• Gouki Minegishi
• Yusuke Iwasawa
• Yutaka Matsuo

Grokking has been actively explored to reveal the mystery of delayed generalization and identifying interpretable representations and algorithms inside the grokked models is a suggestive hint to understanding its mechanism. Grokking on modular addition has been known to implement Fourier representation and its calculation circuits with trigonometric identities in Transformers. Considering the periodicity in modular arithmetic, the natural question is to what extent these explanations and interpretations hold for the grokking on other modular operations beyond addition. For a closer look, we first hypothesize that (1) any modular operations can be characterized with distinctive Fourier representation or internal circuits, (2) grokked models obtain common features transferable among similar operations, and (3) mixing datasets with similar operations promotes grokking. Then, we extensively examine them by learning Transformers on complex modular arithmetic tasks, including polynomials. Our Fourier analysis and novel progress measure for modular arithmetic, Fourier Frequency Density and Fourier Coefficient Ratio, characterize distinctive internal representations of grokked models per modular operation; for instance, polynomials often result in the superposition of the Fourier components seen in elementary arithmetic, but clear patterns do not emerge in challenging non-factorizable polynomials. In contrast, our ablation study on the pre-grokked models reveals that the transferability among the models grokked with each operation can be only limited to specific combinations, such as from elementary arithmetic to linear expressions. Moreover, some multi-task mixtures may lead to co-grokking -- where grokking simultaneously happens for all the tasks -- and accelerate generalization, while others may not find optimal solutions. We empirically provide significant steps towards the interpretability of internal circuits learned through modular operations, where analytical solutions are not attainable.

ICLR Conference 2023 Conference Paper

A System for Morphology-Task Generalization via Unified Representation and Behavior Distillation

• Hiroki Furuta
• Yusuke Iwasawa
• Yutaka Matsuo
• Shixiang Gu

The rise of generalist large-scale models in natural language and vision has made us expect that a massive data-driven approach could achieve broader generalization in other domains such as continuous control. In this work, we explore a method for learning a single policy that manipulates various forms of agents to solve various tasks by distilling a large amount of proficient behavioral data. In order to align input-output (IO) interface among multiple tasks and diverse agent morphologies while preserving essential 3D geometric relations, we introduce morphology-task graph, which treats observations, actions and goals/task in a unified graph representation. We also develop MxT-Bench for fast large-scale behavior generation, which supports procedural generation of diverse morphology-task combinations with a minimal blueprint and hardware-accelerated simulator. Through efficient representation and architecture selection on MxT-Bench, we find out that a morphology-task graph representation coupled with Transformer architecture improves the multi-task performances compared to other baselines including recent discrete tokenization, and provides better prior knowledge for zero-shot transfer or sample efficiency in downstream multi-task imitation learning. Our work suggests large diverse offline datasets, unified IO representation, and policy representation and architecture selection through supervised learning form a promising approach for studying and advancing morphology-task generalization.

NeurIPS Conference 2023 Conference Paper

DreamSparse: Escaping from Plato’s Cave with 2D Diffusion Model Given Sparse Views

• Paul Yoo
• Jiaxian Guo
• Yutaka Matsuo
• Shixiang (Shane) Gu

Synthesizing novel view images from a few views is a challenging but practical problem. Existing methods often struggle with producing high-quality results or necessitate per-object optimization in such few-view settings due to the insufficient information provided. In this work, we explore leveraging the strong 2D priors in pre-trained diffusion models for synthesizing novel view images. 2D diffusion models, nevertheless, lack 3D awareness, leading to distorted image synthesis and compromising the identity. To address these problems, we propose $\textit{DreamSparse}$, a framework that enables the frozen pre-trained diffusion model to generate geometry and identity-consistent novel view images. Specifically, DreamSparse incorporates a geometry module designed to capture features about spatial information from sparse views as a 3D prior. Subsequently, a spatial guidance model is introduced to convert rendered feature maps as spatial information for the generative process. This information is then used to guide the pre-trained diffusion model toencourage the synthesis of geometrically consistent images without further tuning. Leveraging the strong image priors in the pre-trained diffusion models, DreamSparse is capable of synthesizing high-quality novel views for both object and object-centric scene-level images and generalising to open-set images. Experimental results demonstrate that our framework can effectively synthesize novel view images from sparse views and outperforms baselines in both trained and open-set category images. More results can be found on our project page: https: //sites. google. com/view/dreamsparse-webpage.

ICML Conference 2023 Conference Paper

End-to-end Training of Deep Boltzmann Machines by Unbiased Contrastive Divergence with Local Mode Initialization

• Shohei Taniguchi
• Masahiro Suzuki 0001
• Yusuke Iwasawa
• Yutaka Matsuo

We address the problem of biased gradient estimation in deep Boltzmann machines (DBMs). The existing method to obtain an unbiased estimator uses a maximal coupling based on a Gibbs sampler, but when the state is high-dimensional, it takes a long time to converge. In this study, we propose to use a coupling based on the Metropolis-Hastings (MH) and to initialize the state around a local mode of the target distribution. Because of the propensity of MH to reject proposals, the coupling tends to converge in only one step with a high probability, leading to high efficiency. We find that our method allows DBMs to be trained in an end-to-end fashion without greedy pretraining. We also propose some practical techniques to further improve the performance of DBMs. We empirically demonstrate that our training algorithm enables DBMs to show comparable generative performance to other deep generative models, achieving the FID score of 10. 33 for MNIST.

ICLR Conference 2023 Conference Paper

Interaction-Based Disentanglement of Entities for Object-Centric World Models

• Akihiro Nakano
• Masahiro Suzuki 0001
• Yutaka Matsuo

Perceiving the world compositionally in terms of space and time is essential to understanding object dynamics and solving downstream tasks. Object-centric learning using generative models has improved in its ability to learn distinct representations of individual objects and predict their interactions, and how to utilize the learned representations to solve untrained, downstream tasks is a focal question. However, as models struggle to predict object interactions and track the objects accurately, especially for unseen configurations, using object-centric representations in downstream tasks is still a challenge. This paper proposes STEDIE, a new model that disentangles object representations, based on interactions, into interaction-relevant relational features and interaction-irrelevant global features without supervision. Empirical evaluation shows that the proposed model factorizes global features, unaffected by interactions from relational features that are necessary to predict outcome of interactions. We also show that STEDIE achieves better performance in planning tasks and understanding causal relationships. In both tasks, our model not only achieves better performance in terms of reconstruction ability but also utilizes the disentangled representations to solve the tasks in a structured manner.

ICML Conference 2023 Conference Paper

Regularization and Variance-Weighted Regression Achieves Minimax Optimality in Linear MDPs: Theory and Practice

• Toshinori Kitamura
• Tadashi Kozuno
• Yunhao Tang
• Nino Vieillard
• Michal Valko
• Wenhao Yang
• Jincheng Mei
• Pierre Ménard

Mirror descent value iteration (MDVI), an abstraction of Kullback-Leibler (KL) and entropy-regularized reinforcement learning (RL), has served as the basis for recent high-performing practical RL algorithms. However, despite the use of function approximation in practice, the theoretical understanding of MDVI has been limited to tabular Markov decision processes (MDPs). We study MDVI with linear function approximation through its sample complexity required to identify an $\varepsilon$-optimal policy with probability $1-\delta$ under the settings of an infinite-horizon linear MDP, generative model, and G-optimal design. We demonstrate that least-squares regression weighted by the variance of an estimated optimal value function of the next state is crucial to achieving minimax optimality. Based on this observation, we present Variance-Weighted Least-Squares MDVI (VWLS-MDVI), the first theoretical algorithm that achieves nearly minimax optimal sample complexity for infinite-horizon linear MDPs. Furthermore, we propose a practical VWLS algorithm for value-based deep RL, Deep Variance Weighting (DVW). Our experiments demonstrate that DVW improves the performance of popular value-based deep RL algorithms on a set of MinAtar benchmarks.

ICLR Conference 2022 Conference Paper

Generalized Decision Transformer for Offline Hindsight Information Matching

• Hiroki Furuta
• Yutaka Matsuo
• Shixiang Gu

How to extract as much learning signal from each trajectory data has been a key problem in reinforcement learning (RL), where sample inefficiency has posed serious challenges for practical applications. Recent works have shown that using expressive policy function approximators and conditioning on future trajectory information -- such as future states in hindsight experience replay (HER) or returns-to-go in Decision Transformer (DT) -- enables efficient learning of multi-task policies, where at times online RL is fully replaced by offline behavioral cloning (BC), e.g. sequence modeling. We demonstrate that all these approaches are doing hindsight information matching (HIM) -- training policies that can output the rest of trajectory that matches some statistics of future state information. We present Generalized Decision Transformer (GDT) for solving any HIM problem, and show how different choices for the feature function and the anti-causal aggregator not only recover DT as a special case, but also lead to novel Categorical DT (CDT) and Bi-directional DT (BDT) for matching different statistics of the future. For evaluating CDT and BDT, we define offline multi-task state-marginal matching (SMM) and imitation learning (IL) as two generic HIM problems, propose a Wasserstein distance loss as a metric for both, and empirically study them on MuJoCo continuous control benchmarks. Categorical DT, which simply replaces anti-causal summation with anti-causal binning in DT, enables arguably the first effective offline multi-task SMM algorithm that generalizes well to unseen (and even synthetic) multi-modal reward or state-feature distributions. Bi-directional DT, which uses an anti-causal second transformer as the aggregator, can learn to model any statistics of the future and outperforms DT variants in offline multi-task IL, i.e. one-shot IL. Our generalized formulations from HIM and GDT greatly expand the role of powerful sequence modeling architectures in modern RL.

NeurIPS Conference 2022 Conference Paper

Langevin Autoencoders for Learning Deep Latent Variable Models

• Shohei Taniguchi
• Yusuke Iwasawa
• Wataru Kumagai
• Yutaka Matsuo

Markov chain Monte Carlo (MCMC), such as Langevin dynamics, is valid for approximating intractable distributions. However, its usage is limited in the context of deep latent variable models owing to costly datapoint-wise sampling iterations and slow convergence. This paper proposes the amortized Langevin dynamics (ALD), wherein datapoint-wise MCMC iterations are entirely replaced with updates of an encoder that maps observations into latent variables. This amortization enables efficient posterior sampling without datapoint-wise iterations. Despite its efficiency, we prove that ALD is valid as an MCMC algorithm, whose Markov chain has the target posterior as a stationary distribution under mild assumptions. Based on the ALD, we also present a new deep latent variable model named the Langevin autoencoder (LAE). Interestingly, the LAE can be implemented by slightly modifying the traditional autoencoder. Using multiple synthetic datasets, we first validate that ALD can properly obtain samples from target posteriors. We also evaluate the LAE on the image generation task, and show that our LAE can outperform existing methods based on variational inference, such as the variational autoencoder, and other MCMC-based methods in terms of the test likelihood.

NeurIPS Conference 2022 Conference Paper

Large Language Models are Zero-Shot Reasoners

• Takeshi Kojima
• Shixiang (Shane) Gu
• Machel Reid
• Yutaka Matsuo
• Yusuke Iwasawa

Pretrained large language models (LLMs) are widely used in many sub-fields of natural language processing (NLP) and generally known as excellent few-shot learners with task-specific exemplars. Notably, chain of thought (CoT) prompting, a recent technique for eliciting complex multi-step reasoning through step-by-step answer examples, achieved the state-of-the-art performances in arithmetics and symbolic reasoning, difficult system-2 tasks that do not follow the standard scaling laws for LLMs. While these successes are often attributed to LLMs' ability for few-shot learning, we show that LLMs are decent zero-shot reasoners by simply adding ``Let's think step by step'' before each answer. Experimental results demonstrate that our Zero-shot-CoT, using the same single prompt template, significantly outperforms zero-shot LLM performances on diverse benchmark reasoning tasks including arithmetics (MultiArith, GSM8K, AQUA-RAT, SVAMP), symbolic reasoning (Last Letter, Coin Flip), and other logical reasoning tasks (Date Understanding, Tracking Shuffled Objects), without any hand-crafted few-shot examples, e. g. increasing the accuracy on MultiArith from 17. 7% to 78. 7% and GSM8K from 10. 4% to 40. 7% with large-scale InstructGPT model (text-davinci-002), as well as similar magnitudes of improvements with another off-the-shelf large model, 540B parameter PaLM. The versatility of this single prompt across very diverse reasoning tasks hints at untapped and understudied fundamental zero-shot capabilities of LLMs, suggesting high-level, multi-task broad cognitive capabilities may be extracted by simple prompting. We hope our work not only serves as the minimal strongest zero-shot baseline for the challenging reasoning benchmarks, but also highlights the importance of carefully exploring and analyzing the enormous zero-shot knowledge hidden inside LLMs before crafting finetuning datasets or few-shot exemplars.

IJCAI Conference 2022 Conference Paper

Robustifying Vision Transformer without Retraining from Scratch by Test-Time Class-Conditional Feature Alignment

• Takeshi Kojima
• Yutaka Matsuo
• Yusuke Iwasawa

Vision Transformer (ViT) is becoming more popular in image processing. Specifically, we investigate the effectiveness of test-time adaptation (TTA) on ViT, a technique that has emerged to correct its prediction during test-time by itself. First, we benchmark various test-time adaptation approaches on ViT-B16 and ViT-L16. It is shown that the TTA is effective on ViT and the prior-convention (sensibly selecting modulation parameters) is not necessary when using proper loss function. Based on the observation, we propose a new test-time adaptation method called class-conditional feature alignment (CFA), which minimizes both the class-conditional distribution differences and the whole distribution differences of the hidden representation between the source and target in an online manner. Experiments of image classification tasks on common corruption (CIFAR-10-C, CIFAR-100-C, and ImageNet-C) and domain adaptation (digits datasets and ImageNet-Sketch) show that CFA stably outperforms the existing baselines on various datasets. We also verify that CFA is model agnostic by experimenting on ResNet, MLP-Mixer, and several ViT variants (ViT-AugReg, DeiT, and BeiT). Using BeiT backbone, CFA achieves 19. 8% top-1 error rate on ImageNet-C, outperforming the existing test-time adaptation baseline 44. 0%. This is a state-of-the-art result among TTA methods that do not need to alter training phase.

NeurIPS Conference 2021 Conference Paper

Co-Adaptation of Algorithmic and Implementational Innovations in Inference-based Deep Reinforcement Learning

• Hiroki Furuta
• Tadashi Kozuno
• Tatsuya Matsushima
• Yutaka Matsuo
• Shixiang (Shane) Gu

Recently many algorithms were devised for reinforcement learning (RL) with function approximation. While they have clear algorithmic distinctions, they also have many implementation differences that are algorithm-independent and sometimes under-emphasized. Such mixing of algorithmic novelty and implementation craftsmanship makes rigorous analyses of the sources of performance improvements across algorithms difficult. In this work, we focus on a series of off-policy inference-based actor-critic algorithms -- MPO, AWR, and SAC -- to decouple their algorithmic innovations and implementation decisions. We present unified derivations through a single control-as-inference objective, where we can categorize each algorithm as based on either Expectation-Maximization (EM) or direct Kullback-Leibler (KL) divergence minimization and treat the rest of specifications as implementation details. We performed extensive ablation studies, and identified substantial performance drops whenever implementation details are mismatched for algorithmic choices. These results show which implementation or code details are co-adapted and co-evolved with algorithms, and which are transferable across algorithms: as examples, we identified that tanh Gaussian policy and network sizes are highly adapted to algorithmic types, while layer normalization and ELU are critical for MPO's performances but also transfer to noticeable gains in SAC. We hope our work can inspire future work to further demystify sources of performance improvements across multiple algorithms and allow researchers to build on one another's both algorithmic and implementational innovations.

ICLR Conference 2021 Conference Paper

Deployment-Efficient Reinforcement Learning via Model-Based Offline Optimization

• Tatsuya Matsushima
• Hiroki Furuta
• Yutaka Matsuo
• Ofir Nachum
• Shixiang Gu

Most reinforcement learning (RL) algorithms assume online access to the environment, in which one may readily interleave updates to the policy with experience collection using that policy. However, in many real-world applications such as health, education, dialogue agents, and robotics, the cost or potential risk of deploying a new data-collection policy is high, to the point that it can become prohibitive to update the data-collection policy more than a few times during learning. With this view, we propose a novel concept of deployment efficiency, measuring the number of distinct data-collection policies that are used during policy learning. We observe that naïvely applying existing model-free offline RL algorithms recursively does not lead to a practical deployment-efficient and sample-efficient algorithm. We propose a novel model-based algorithm, Behavior-Regularized Model-ENsemble (BREMEN), that not only performs better than or comparably as the state-of-the-art dynamic-programming-based and concurrently-proposed model-based offline approaches on existing benchmarks, but can also effectively optimize a policy offline using 10-20 times fewer data than prior works. Furthermore, the recursive application of BREMEN achieves impressive deployment efficiency while maintaining the same or better sample efficiency, learning successful policies from scratch on simulated robotic environments with only 5-10 deployments, compared to typical values of hundreds to millions in standard RL baselines.

ICLR Conference 2021 Conference Paper

Group Equivariant Conditional Neural Processes

• Makoto Kawano
• Wataru Kumagai
• Akiyoshi Sannai
• Yusuke Iwasawa
• Yutaka Matsuo

We present the group equivariant conditional neural process (EquivCNP), a meta-learning method with permutation invariance in a data set as in conventional conditional neural processes (CNPs), and it also has transformation equivariance in data space. Incorporating group equivariance, such as rotation and scaling equivariance, provides a way to consider the symmetry of real-world data. We give a decomposition theorem for permutation-invariant and group-equivariant maps, which leads us to construct EquivCNPs with an infinite-dimensional latent space to handle group symmetries. In this paper, we build architecture using Lie group convolutional layers for practical implementation. We show that EquivCNP with translation equivariance achieves comparable performance to conventional CNPs in a 1D regression task. Moreover, we demonstrate that incorporating an appropriate Lie group equivariance, EquivCNP is capable of zero-shot generalization for an image-completion task by selecting an appropriate Lie group equivariance.

ICML Conference 2021 Conference Paper

Policy Information Capacity: Information-Theoretic Measure for Task Complexity in Deep Reinforcement Learning

• Hiroki Furuta
• Tatsuya Matsushima
• Tadashi Kozuno
• Yutaka Matsuo
• Sergey Levine
• Ofir Nachum
• Shixiang Gu

Progress in deep reinforcement learning (RL) research is largely enabled by benchmark task environments. However, analyzing the nature of those environments is often overlooked. In particular, we still do not have agreeable ways to measure the difficulty or solvability of a task, given that each has fundamentally different actions, observations, dynamics, rewards, and can be tackled with diverse RL algorithms. In this work, we propose policy information capacity (PIC) – the mutual information between policy parameters and episodic return – and policy-optimal information capacity (POIC) – between policy parameters and episodic optimality – as two environment-agnostic, algorithm-agnostic quantitative metrics for task difficulty. Evaluating our metrics across toy environments as well as continuous control benchmark tasks from OpenAI Gym and DeepMind Control Suite, we empirically demonstrate that these information-theoretic metrics have higher correlations with normalized task solvability scores than a variety of alternatives. Lastly, we show that these metrics can also be used for fast and compute-efficient optimizations of key design parameters such as reward shaping, policy architectures, and MDP properties for better solvability by RL algorithms without ever running full RL experiments.

NeurIPS Conference 2021 Conference Paper

Test-Time Classifier Adjustment Module for Model-Agnostic Domain Generalization

• Yusuke Iwasawa
• Yutaka Matsuo

This paper presents a new algorithm for domain generalization (DG), \textit{test-time template adjuster (T3A)}, aiming to robustify a model to unknown distribution shift. Unlike existing methods that focus on \textit{training phase}, our method focuses \textit{test phase}, i. e. , correcting its prediction by itself during test time. Specifically, T3A adjusts a trained linear classifier (the last layer of deep neural networks) with the following procedure: (1) compute a pseudo-prototype representation for each class using online unlabeled data augmented by the base classifier trained in the source domains, (2) and then classify each sample based on its distance to the pseudo-prototypes. T3A is back-propagation-free and modifies only the linear layer; therefore, the increase in computational cost during inference is negligible and avoids the catastrophic failure might caused by stochastic optimization. Despite its simplicity, T3A can leverage knowledge about the target domain by using off-the-shelf test-time data and improve performance. We tested our method on four domain generalization benchmarks, namely PACS, VLCS, OfficeHome, and TerraIncognita, along with various backbone networks including ResNet18, ResNet50, Big Transfer (BiT), Vision Transformers (ViT), and MLP-Mixer. The results show T3A stably improves performance on unseen domains across choices of backbone networks, and outperforms existing domain generalization methods.

AAAI Conference 2021 Conference Paper

Variational Inference for Learning Representations of Natural Language Edits

• Edison Marrese-Taylor
• Machel Reid
• Yutaka Matsuo

Document editing has become a pervasive component of the production of information, with version control systems enabling edits to be efficiently stored and applied. In light of this, the task of learning distributed representations of edits has been recently proposed. With this in mind, we propose a novel approach that employs variational inference to learn a continuous latent space of vector representations to capture the underlying semantic information with regard to the document editing process. We achieve this by introducing a latent variable to explicitly model the aforementioned features. This latent variable is then combined with a document representation to guide the generation of an edited version of this document. Additionally, to facilitate standardized automatic evaluation of edit representations, which has heavily relied on direct human input thus far, we also propose a suite of downstream tasks, PEER, specifically designed to measure the quality of edit representations in the context of natural language processing.

IJCAI Conference 2020 Conference Paper

Stabilizing Adversarial Invariance Induction from Divergence Minimization Perspective

• Yusuke Iwasawa
• Kei Akuzawa
• Yutaka Matsuo

Adversarial invariance induction (AII) is a generic and powerful framework for enforcing an invariance to nuisance attributes into neural network representations. However, its optimization is often unstable and little is known about its practical behavior. This paper presents an analysis of the reasons for the optimization difficulties and provides a better optimization procedure by rethinking AII from a divergence minimization perspective. Interestingly, this perspective indicates a cause of the optimization difficulties: it does not ensure proper divergence minimization, which is a requirement of the invariant representations. We then propose a simple variant of AII, called invariance induction by discriminator matching, which takes into account the divergence minimization interpretation of the invariant representations. Our method consistently achieves near-optimal invariance in toy datasets with various configurations in which the original AII is catastrophically unstable. Extentive experiments on four real-world datasets also support the superior performance of the proposed method, leading to improved user anonymization and domain generalization.

IJCAI Conference 2017 Conference Paper

Privacy Issues Regarding the Application of DNNs to Activity-Recognition using Wearables and Its Countermeasures by Use of Adversarial Training

• Yusuke Iwasawa
• Kotaro Nakayama
• Ikuko Yairi
• Yutaka Matsuo

Deep neural networks have been successfully applied to activity recognition with wearables in terms of recognition performance. However, the black-box nature of neural networks could lead to privacy concerns. Namely, generally it is hard to expect what neural networks learn from data, and so they possibly learn features that highly discriminate user-information unintentionally, which increases the risk of information-disclosure. In this study, we analyzed the features learned by conventional deep neural networks when applied to data of wearables to confirm this phenomenon. Based on the results of our analysis, we propose the use of an adversarial training framework to suppress the risk of sensitive/unintended information disclosure. Our proposed model considers both an adversarial user classifier and a regular activity-classifier during training, which allows the model to learn representations that help the classifier to distinguish the activities but which, at the same time, prevents it from accessing user-discriminative information. This paper provides an empirical validation of the privacy issue and efficacy of the proposed method using three activity recognition tasks based on data of wearables. The empirical validation shows that our proposed method suppresses the concerns without any significant performance degradation, compared to conventional deep nets on all three tasks.

IJCAI Conference 2011 Conference Paper

Mining Longitudinal Network for Predicting Company Value

• Yingzi Jin
• Ching-Yung Lin
• Yutaka Matsuo
• Mitsuru Ishizuka

Real-world social networks are dynamic in nature. Companies continue to collaborate, align strategically, acquire, and merge over time, and receive positive/negative impact from other companies. Consequently, their performance changes with time. If one can understand what types of network changes affect a company's value, he/she can predict the future value of the company, grasp industry innovations, and make business more successful. However, it often requires continuous records of relational changes, which are often difficult to track for companies, and the models of mining longitudinal network are quite complicated. In this study, we developed algorithms and a system to infer large-scale evolutionary company networks from public news during 1981--2009. Then, based on how networks change over time, as well as the financial information of the companies, we predicted company profit growth. This is the first study of longitudinal network-mining-based company performance analysis in the literature.

IJCAI Conference 2011 Conference Paper

Relation Adaptation: Learning to Extract Novel Relations with Minimum Supervision

• Danushka Bollegala
• Yutaka Matsuo
• Mitsuru Ishizuka

Extracting the relations that exist between two entities is an important step in numerousWeb-related tasks such as information extraction. A supervised relation extraction system that is trained to extract a particular relation type might not accurately extract a new type of a relation for which it has not been trained. However, it is costly to create training data manually for every new relation type that one might want to extract. We propose a method to adapt an existing relation extraction system to extractnew relation types with minimum supervision. Our proposed method comprises two stages: learning a lower-dimensional projection between different relations, and learning a relational classifier for the target relation type with instance sampling. We evaluate the proposed method using a dataset that contains 2000 instances for 20 different relation types. Our experimental results show that the proposed method achieves a statistically significant macro-average F-score of 62. 77. Moreover, the proposed method outperforms numerous baselines and a previously proposed weakly-supervised relation extraction method.

ECAI Conference 2008 Conference Paper

WWW sits the SAT: Measuring Relational Similarity on the Web

• Danushka Bollegala
• Yutaka Matsuo
• Mitsuru Ishizuka

Measuring relational similarity between words is important in numerous natural language processing tasks such as solving analogy questions and classifying noun-modifier relations. We propose a method to measure the similarity between semantic relations that hold between two pairs of words using a web search engine. First, each pair of words is represented by a vector of automatically extracted lexical patterns. Then a Support Vector Machine is trained to recognize word pairs with similar semantic relations. We evaluate the proposed method on SAT multiple-choice word-analogy questions. The proposed method achieves a score of 40% which is comparable with relational similarity measures which use manually created resources such as WordNet. The proposed method significantly reduces the time taken by previously proposed computationally intensive methods, such as latent relational analysis, to process 374 analogy questions from 8 days to less than 6 hours.

IJCAI Conference 2007 Conference Paper

• Junichiro Mori
• Yutaka Matsuo
• Mitsuru Ishizuka

Social networks have recently garnered considerable interest. With the intention of utilizing social networks for the Semantic Web, several studies have examined automatic extraction of social networks. However, most methods have addressed extraction of the strength of relations. Our goal is extracting the underlying relations between entities that are embedded in social networks. To this end, we propose a method that automatically extracts labels that describe relations among entities. Fundamentally, the method clusters similar entity pairs according to their collective contexts in Web documents. The descriptive labels for relations are obtained from results of clustering. The proposed method is entirely unsupervised and is easily incorporated with existing social network extraction methods. Our experiments conducted on entities in researcher social networks and political social networks achieved clustering with high precision and recall. The results showed that our method is able to extract appropriate relation labels to represent relations among entities in the social networks.

IJCAI Conference 2007 Conference Paper

• Yutaka Matsuo
• Naoaki Okazaki
• Kiyoshi Izumi
• Yoshiyuki Nakamura
• Takuichi Nishimura
• Koiti Hasida
• Hideyuki Nakashima

Recent development of location technologies enables us to obtain the location history of users. paper proposes a new method to infer users' longterm properties from their respective location histories. Counting the instances of sensor detection every user, we can obtain a sensor-user matrix. After generating features from the matrix, a machine learning approach is taken to automatically users into different categories for each user property. Inspired by information retrieval research, problem to infer user properties is reduced categorization problem. We compare weightings of several features and also propose sensor weighting. Our algorithms are evaluated using experimental location data in an office environment.

AAMAS Conference 2007 Conference Paper

Emergence of Global Network Property based on Multi-agent Voting Model

• Kosuke Shinoda
• Yutaka Matsuo
• Hideyuki Nakashima

Recent studies have shown that various models can explain the emergence of complex networks, such as scale-free and small-world networks. This paper presents a different model to generate complex networks using a multi-agent approach. Each node is considered as an agent. Based on voting by all agents, edges are added repeatedly. We use four different kinds of centrality measures as a utility functions for agents. Depending on the centrality measure, the resultant networks differ considerably: typically, closeness centrality generates a scale-free network, degree centrality produces a random graph, betweenness centrality favors a regular graph, and eigenvector centrality brings a complete subgraph. The importance of the network structure among agents is widely noted in the multi-agent research literature. This paper contributes new insights into the connection between agents' local behavior and the global property of the network structure. We describe a detailed analysis on why these structures emerge, and present a discussion of the possible expansion and application of the model.

AAAI Conference 2007 Conference Paper

Robust Estimation of Google Counts for Social Network Extraction

• Yutaka Matsuo

Various studies within NLP and Semantic Web use the so-called Google count, which is the hit count on a query returned by a search engine (not only Google). However, sometimes the Google count is unreliable, especially when the count is large, or when advanced operators such as OR and NOT are used. In this paper, we propose a novel algorithm that estimates the Google count robustly. It (i) uses the co-occurrence of terms as evidence to estimate the occurrence of a given word, and (ii) integrates multiple evidence for robust estimation. We evaluated our algorithm for more than 2000 queries on three datasets using Google, Yahoo! and MSN search engine. Our algorithm also provides estimate counts for any classifier that judges a web page as positive or negative. Consequently, we can estimate the number of documents with included references of a particular person (among namesakes) on the entire web.

AAMAS Conference 2007 Conference Paper

Socially Embedded Multi Agent Based Simulation of Financial Market

• Kiyoshi Izumi
• Hiroki Matsui
• Yutaka Matsuo

This paper proposed a new approach that integrated an artificial market simulation and text-mining with real information. In this approach, economic trends were extracted from text data circulating in the real world. Then, the trends were inputted into the market simulation. The simulation could support users' action to the actual market. This approach was used for the decision of exchange rate policy and suggested that the operation by intervention was effective for the stabilization of the yen-dollar rate in 1995. Our simulation revealed that the action rule proposed by our system could reduce over 70% of rate fluctuation. This approach can offer a useful social simulation as a tool to users.

ECAI Conference 2006 Conference Paper

Disambiguating Personal Names on the Web Using Automatically Extracted Key Phrases

• Danushka Bollegala
• Yutaka Matsuo
• Mitsuru Ishizuka

When you search for information regarding a particular person on the web, a search engine returns many pages. Some of these pages may be for people with the same name. How can we disambiguate these different people with the same name? This paper presents an unsupervised algorithm which produces unique phrases to disambiguate different people with the same name (i. e. namesakes). Our algorithm takes in a personal name and outputs multiple sets of phrases which uniquely identify the different namesakes on the web. These phrases could then be added to the query to narrow down the search to a specific namesake. We evaluated the algorithm on a collection of documents retreived from the Web. Experimental results show a significant improvement over the existing methods proposed for this task.

AAAI Conference 2006 Conference Paper

Spinning Multiple Social Networks for Semantic Web

• Yutaka Matsuo
• Hideaki Takeda
• Danushka Bollegara
• Takuichi Nishimura

Social networks are important for the Semantic Web. Several means can be used to obtain social networks: using social networking services, aggregating Friendof-a-Friend (FOAF) documents, mining text information on the Web or in e-mail messages, and observing face-to-face communication using sensors. Integrating multiple social networks is a key issue for further utilization of social networks in the Semantic Web. This paper describes our attempt to extract, analyze and integrate multiple social networks from the same community: user-registered knows networks, web-mined collaborator networks, and face-to-face meets networks. We operated a social network-based community support system called Polyphonet at the 17th, 18th and 19th Annual Conferences of the Japan Society of Artificial Intelligence (JSAI2003, JSAI2004, and JSAI2005) and at The International Conference on Ubiquitous Computing (UbiComp 2005). Multiple social networks were obtained and analyzed. We discuss the integration of multiple networks based on the analyses.

ECAI Conference 2004 Conference Paper

Finding Social Network for Trust Calculation

• Yutaka Matsuo
• Hironori Tomobe
• Kôiti Hasida
• Mitsuru Ishizuka

IROS Conference 2004 Conference Paper

Smart navigation by lightweight components coordination in ubiquitous computing

• Shigeyoshi Hiratsuka
• Yutaka Matsuo
• Akio Sashima
• Akira Takagi
• Noriaki Izumi
• Koichi Kurumatani

We discuss a smart navigation system by route planning which makes active use of spatial semantics. The central concept of the system is coordination of lightweight components as services, such as business process modeling. Our proposed navigation system consists of three main components: a route planner, an object repository, and a space functionality retriever (SFR). The route planner is the component for inquisitions. The object repository deals with real-world information. Then, the SFR deals with knowledge about space represented by XML-based descriptions. Using these components, we realize a smart navigation system that enables a route planner to suggest alternate solutions to end-users. Our final target is to develop advanced location-based information services that surpass conventional ones, thereby realizing an intelligent space.