Searcharxiv⌕ Search

arXiv subjects

Takayuki Osogami

Publications and source records attributed to Takayuki Osogami.

At least 19 recordsLinked to original sources
arXiv2026

Distorted Distributional Policy Evaluation for Offline Reinforcement Learning

While Distributional Reinforcement Learning (DRL) methods have demonstrated strong performance in online settings, its success in offline scenarios remains limited. We hypothesize that a key limitation of existing offline DRL methods lies in their approach to uniformly underestimate return quantiles. This uniform pessimism can lead to overly conservative value estimates, ultimately hindering generalization and performance. To address this, we introduce a novel concept called quantile distortion, which enables non-uniform pessimism by adjusting the degree of conservatism based on the availability of supporting data. Our approach is grounded in theoretical analysis and empirically validated, demonstrating improved performance over uniform pessimism.

cs.LG↗
arXiv2025

AI Agents Should be Regulated Based on the Extent of Their Autonomous Operations

This position paper argues that AI agents should be regulated by the extent to which they operate autonomously. AI agents with long-term planning and strategic capabilities can pose significant risks of human extinction and irreversible global catastrophes. While existing regulations often focus on computational scale as a proxy for potential harm, we argue that such measures are insufficient for assessing the risks posed by agents whose capabilities arise primarily from inference-time computation. To support our position, we discuss relevant regulations and recommendations from scientists regarding existential risks, as well as the advantages of using action sequences -- which reflect the degree of an agent's autonomy -- as a more suitable measure of potential impact than existing metrics that rely on observing environmental states.

cs.CY↗
arXiv2025

Achieving PAC Guarantees in Mechanism Design through Multi-Armed Bandits

We analytically derive a class of optimal solutions to a linear program (LP) for automated mechanism design that satisfies efficiency, incentive compatibility, strong budget balance (SBB), and individual rationality (IR), where SBB and IR are enforced in expectation. These solutions can be expressed using a set of essential variables whose cardinality is exponentially smaller than the total number of variables in the original formulation. However, evaluating a key term in the solutions requires exponentially many optimization steps as the number of players $N$ increases. We address this by translating the evaluation of this term into a multi-armed bandit (MAB) problem and develop a probably approximately correct (PAC) estimator with asymptotically optimal sample complexity. This MAB-based approach reduces the optimization complexity from exponential to $O(N\log N)$. Numerical experiments confirm that our method efficiently computes mechanisms with the target properties, scaling to problems with up to $N=128$ players -- substantially improving over prior work.

cs.GT↗
arXiv2024

Socially efficient mechanism on the minimum budget

In social decision-making among strategic agents, a universal focus lies on the balance between social and individual interests. Socially efficient mechanisms are thus desirably designed to not only maximize the social welfare but also incentivize the agents for their own profit. Under a generalized model that includes applications such as double auctions and trading networks, this study establishes a socially efficient (SE), dominant-strategy incentive compatible (DSIC), and individually rational (IR) mechanism with the minimum total budget expensed to the agents. The present method exploits discrete and known type domains to reduce a set of constraints into the shortest path problem in a weighted graph. In addition to theoretical derivation, we substantiate the optimality of the proposed mechanism through numerical experiments, where it certifies strictly lower budget than Vickery-Clarke-Groves (VCG) mechanisms for a wide class of instances.

cs.GT↗
arXiv2023

Who Benefits from a Multi-Cloud Market? A Trading Networks Based Analysis

In enterprise cloud computing, there is a big and increasing investment to move to multi-cloud computing, which allows enterprises to seamlessly utilize IT resources from multiple cloud providers, so as to take advantage of different cloud providers' capabilities and costs. This investment raises several key questions: Will multi-cloud always be more beneficial to the cloud users? How will this impact the cloud providers? Is it possible to create a multi-cloud market that is beneficial to all participants? In this work, we begin addressing these questions by using the game theoretic model of trading networks and formally compare between the single and multi-cloud markets. This comparson a) provides a sufficient condition under which the multi-cloud network can be considered more efficient than the single cloud one in the sense that a centralized coordinator having full information can impose an outcome that is strongly Pareto-dominant for all players and b) shows a surprising result that without centralized coordination, settings are possible in which even the cloud buyers' utilities may decrease when moving from a single cloud to a multi-cloud network. As these two results emphasize the need for centralized coordination to ensure a Pareto-dominant outcome and as the aforementioned Pareto-dominant result requires truthful revelation of participant's private information, we provide an automated mechanism design (AMD) approach, which, in the Bayesian setting, finds mechanisms which result in expectation in such Pareto-dominant outcomes, and in which truthful revelation of the parties' private information is the dominant strategy. We also provide empirical analysis to show the validity of our AMD approach.

cs.GT↗
arXiv2023

Regression with Sensor Data Containing Incomplete Observations

This paper addresses a regression problem in which output label values are the results of sensing the magnitude of a phenomenon. A low value of such labels can mean either that the actual magnitude of the phenomenon was low or that the sensor made an incomplete observation. This leads to a bias toward lower values in labels and the resultant learning because labels may have lower values due to incomplete observations, even if the actual magnitude of the phenomenon was high. Moreover, because an incomplete observation does not provide any tags indicating incompleteness, we cannot eliminate or impute them. To address this issue, we propose a learning algorithm that explicitly models incomplete observations corrupted with an asymmetric noise that always has a negative value. We show that our algorithm is unbiased as if it were learned from uncorrupted data that does not involve incomplete observations. We demonstrate the advantages of our algorithm through numerical experiments.

cs.LG↗
arXiv2022

Mechanism Learning for Trading Networks

We study the problem of designing mechanisms for trading networks that satisfy four desired properties: dominant-strategy incentive compatibility, efficiency, weak budget balance (WBB), and individual rationality (IR). Although there exist mechanisms that simultaneously satisfy these properties ex post for combinatorial auctions, we prove the impossibility that such mechanisms do not exist for a broad class of trading networks. We thus propose approaches for computing and learning the mechanisms that satisfy the four properties, in a Bayesian setting, where WBB and IR, respectively, are relaxed to ex ante and interim. For computational and sample efficiency, we introduce several techniques, including game theoretical analysis to reduce the input feature space. We empirically demonstrate that the proposed approaches successfully find the mechanisms with the four properties for those trading networks where the impossibility holds ex post.

cs.GT↗
arXiv2022

Online Learning in Supply-Chain Games

We study a repeated game between a supplier and a retailer who want to maximize their respective profits without full knowledge of the problem parameters. After characterizing the uniqueness of the Stackelberg equilibrium of the stage game with complete information, we show that even with partial knowledge of the joint distribution of demand and production costs, natural learning dynamics guarantee convergence of the joint strategy profile of supplier and retailer to the Stackelberg equilibrium of the stage game. We also prove finite-time bounds on the supplier's regret and asymptotic bounds on the retailer's regret, where the specific rates depend on the type of knowledge preliminarily available to the players. In the special case when the supplier is not strategic (vertical integration), we prove optimal finite-time regret bounds on the retailer's regret (or, equivalently, the social welfare) when costs and demand are adversarially generated and the demand is censored.

cs.GT↗
arXiv2022

Biases in In Silico Evaluation of Molecular Optimization Methods and Bias-Reduced Evaluation Methodology

We are interested in in silico evaluation methodology for molecular optimization methods. Given a sample of molecules and their properties of our interest, we wish not only to train an agent that can find molecules optimized with respect to the target property but also to evaluate its performance. A common practice is to train a predictor of the target property on the sample and use it for both training and evaluating the agent. We show that this evaluator potentially suffers from two biases; one is due to misspecification of the predictor and the other to reusing the same sample for training and evaluation. We discuss bias reduction methods for each of the biases comprehensively, and empirically investigate their effectiveness.

cs.LG↗
arXiv2019

Real-time tree search with pessimistic scenarios

Autonomous agents need to make decisions in a sequential manner, under partially observable environment, and in consideration of how other agents behave. In critical situations, such decisions need to be made in real time for example to avoid collisions and recover to safe conditions. We propose a technique of tree search where a deterministic and pessimistic scenario is used after a specified depth. Because there is no branching with the deterministic scenario, the proposed technique allows us to take into account the events that can occur far ahead in the future. The effectiveness of the proposed technique is demonstrated in Pommerman, a multi-agent environment used in a NeurIPS 2018 competition, where the agents that implement the proposed technique have won the first and third places.

cs.AI↗
arXiv2019

Visual analytics for team-based invasion sports with significant events and Markov reward process

In team-based invasion sports such as soccer and basketball, analytics is important for teams to understand their performance and for audiences to understand matches better. The present work focuses on performing visual analytics to evaluate the value of any kind of event occurring in a sports match with a continuous parameter space. Here, the continuous parameter space involves the time, location, score, and other parameters. Because the spatiotemporal data used in such analytics is a low-level representation and has a very large size, however, traditional analytics may need to discretize the continuous parameter space (e.g., subdivide the playing area) or use a local feature to limit the analysis to specific events (e.g., only shots). These approaches make evaluation impossible for any kind of event with a continuous parameter space. To solve this problem, we consider a whole match as a Markov chain of significant events, so that event values can be estimated with a continuous parameter space by solving the Markov chain with a machine learning model. The significant events are first extracted by considering the time-varying distribution of players to represent the whole match. Then, the extracted events are redefined as different states with the continuous parameter space and built as a Markov chain so that a Markov reward process can be applied. Finally, the Markov reward process is solved by a customized fitted-value iteration algorithm so that the event values with the continuous parameter space can be predicted by a regression model. As a result, the event values can be visually inspected over the whole playing field under arbitrary given conditions. Experimental results with real soccer data show the effectiveness of the proposed system.

cs.AI↗
arXiv2019

Boltzmann machines for time-series

We review Boltzmann machines extended for time-series. These models often have recurrent structure, and back propagration through time (BPTT) is used to learn their parameters. The per-step computational complexity of BPTT in online learning, however, grows linearly with respect to the length of preceding time-series (i.e., learning rule is not local in time), which limits the applicability of BPTT in online learning. We then review dynamic Boltzmann machines (DyBMs), whose learning rule is local in time. DyBM's learning rule relates to spike-timing dependent plasticity (STDP), which has been postulated and experimentally confirmed for biological neural networks.

cs.NE↗
arXiv2019

Boltzmann machines and energy-based models

We review Boltzmann machines and energy-based models. A Boltzmann machine defines a probability distribution over binary-valued patterns. One can learn parameters of a Boltzmann machine via gradient based approaches in a way that log likelihood of data is increased. The gradient and Hessian of a Boltzmann machine admit beautiful mathematical representations, although computing them is in general intractable. This intractability motivates approximate methods, including Gibbs sampler and contrastive divergence, and tractable alternatives, namely energy-based models.

cs.NE↗
arXiv2018

Time-Discounting Convolution for Event Sequences with Ambiguous Timestamps

This paper proposes a method for modeling event sequences with ambiguous timestamps, a time-discounting convolution. Unlike in ordinary time series, time intervals are not constant, small time-shifts have no significant effect, and inputting timestamps or time durations into a model is not effective. The criteria that we require for the modeling are providing robustness against time-shifts or timestamps uncertainty as well as maintaining the essential capabilities of time-series models, i.e., forgetting meaningless past information and handling infinite sequences. The proposed method handles them with a convolutional mechanism across time with specific parameterizations, which efficiently represents the event dependencies in a time-shift invariant manner while discounting the effect of past events, and a dynamic pooling mechanism, which provides robustness against the uncertainty in timestamps and enhances the time-discounting capability by dynamically changing the pooling window size. In our learning algorithm, the decaying and dynamic pooling mechanisms play critical roles in handling infinite and variable length sequences. Numerical experiments on real-world event sequences with ambiguous timestamps and ordinary time series demonstrated the advantages of our method.

cs.LG↗
arXiv2017

Dynamic Boltzmann Machines for Second Order Moments and Generalized Gaussian Distributions

Dynamic Boltzmann Machine (DyBM) has been shown highly efficient to predict time-series data. Gaussian DyBM is a DyBM that assumes the predicted data is generated by a Gaussian distribution whose first-order moment (mean) dynamically changes over time but its second-order moment (variance) is fixed. However, in many financial applications, the assumption is quite limiting in two aspects. First, even when the data follows a Gaussian distribution, its variance may change over time. Such variance is also related to important temporal economic indicators such as the market volatility. Second, financial time-series data often requires learning datasets generated by the generalized Gaussian distribution with an additional shape parameter that is important to approximate heavy-tailed distributions. Addressing those aspects, we show how to extend DyBM that results in significant performance improvement in predicting financial time-series data.

stat.ML↗
arXiv2017

Load Balancing for Skewed Streams on Heterogeneous Cluster

Streaming applications frequently encounter skewed workloads and execute on heterogeneous clusters. Optimal resource utilization in such adverse conditions becomes a challenge, as it requires inferring the resource capacities and input distribution at run time. In this paper, we tackle the aforementioned challenges by modeling them as a load balancing problem. We propose a novel partitioning strategy called Consistent Grouping (CG), which enables each processing element instance (PEI) to process the workload according to its capacity. The main idea behind CG is the notion of small, equal-sized virtual workers at the sources, which are assigned to workers based on their capacities. We provide a theoretical analysis of the proposed algorithm and show via extensive empirical evaluation that our proposed scheme outperforms the state-of-the-art approaches, like key grouping. In particular, CG achieves 3.44x better performance in terms of latency compared to key grouping.

cs.DC↗