Kian Hsiang Low

AAAI Conference 2019 Conference Paper

Collective Online Learning of Gaussian Processes in Massive Multi-Agent Systems

• Trong Nghia Hoang
• Quang Minh Hoang
• Kian Hsiang Low
• Jonathan How

This paper presents a novel Collective Online Learning of Gaussian Processes (COOL-GP) framework for enabling a massive number of GP inference agents to simultaneously perform (a) efficient online updates of their GP models using their local streaming data with varying correlation structures and (b) decentralized fusion of their resulting online GP models with different learned hyperparameter settings and inducing inputs. To realize this, we exploit the notion of a common encoding structure to encapsulate the local streaming data gathered by any GP inference agent into summary statistics based on our proposed representation, which is amenable to both an efficient online update via an importance sampling trick as well as multi-agent model fusion via decentralized message passing that can exploit sparse connectivity among agents for improving efficiency and enhance the robustness of our framework against transmission loss. We provide a rigorous theoretical analysis of the approximation loss arising from our proposed representation to achieve efficient online updates and model fusion. Empirical evaluations show that COOL-GP is highly effective in model fusion, resilient to information disparity between agents, robust to transmission loss, and can scale to thousands of agents.

AAAI Conference 2018 Conference Paper

Decentralized High-Dimensional Bayesian Optimization With Factor Graphs

• Trong Nghia Hoang
• Quang Minh Hoang
• Ruofei Ouyang
• Kian Hsiang Low

This paper presents a novel decentralized high-dimensional Bayesian optimization (DEC-HBO) algorithm that, in contrast to existing HBO algorithms, can exploit the interdependent effects of various input components on the output of the unknown objective function f for boosting the BO performance and still preserve scalability in the number of input dimensions without requiring prior knowledge or the existence of a low (effective) dimension of the input space. To realize this, we propose a sparse yet rich factor graph representation of f to be exploited for designing an acquisition function that can be similarly represented by a sparse factor graph and hence be efficiently optimized in a decentralized manner using distributed message passing. Despite richly characterizing the interdependent effects of the input components on the output of f with a factor graph, DEC-HBO can still guarantee no-regret performance asymptotically. Empirical evaluation on synthetic and real-world experiments (e. g. , sparse Gaussian process model with 1811 hyperparameters) shows that DEC-HBO outperforms the state-of-the-art HBO algorithms.

AAAI Conference 2018 Conference Paper

Gaussian Process Decentralized Data Fusion Meets Transfer Learning in Large-Scale Distributed Cooperative Perception

• Ruofei Ouyang
• Kian Hsiang Low

This paper presents novel Gaussian process decentralized data fusion algorithms exploiting the notion of agent-centric support sets for distributed cooperative perception of largescale environmental phenomena. To overcome the limitations of scale in existing works, our proposed algorithms allow every mobile sensing agent to choose a different support set and dynamically switch to another during execution for encapsulating its own data into a local summary that, perhaps surprisingly, can still be assimilated with the other agents’ local summaries (i. e. , based on their current choices of support sets) into a globally consistent summary to be used for predicting the phenomenon. To achieve this, we propose a novel transfer learning mechanism for a team of agents capable of sharing and transferring information encapsulated in a summary based on a support set to that utilizing a different support set with some loss that can be theoretically bounded and analyzed. To alleviate the issue of information loss accumulating over multiple instances of transfer learning, we propose a new information sharing mechanism to be incorporated into our algorithms in order to achieve memory-efficient lazy transfer learning. Empirical evaluation on real-world datasets show that our algorithms outperform the state-of-the-art methods.

AAAI Conference 2017 Conference Paper

A Generalized Stochastic Variational Bayesian Hyperparameter Learning Framework for Sparse Spectrum Gaussian Process Regression

• Quang Minh Hoang
• Trong Nghia Hoang
• Kian Hsiang Low

While much research effort has been dedicated to scaling up sparse Gaussian process (GP) models based on inducing variables for big data, little attention is afforded to the other less explored class of low-rank GP approximations that exploit the sparse spectral representation of a GP kernel. This paper presents such an effort to advance the state of the art of sparse spectrum GP models to achieve competitive predictive performance for massive datasets. Our generalized framework of stochastic variational Bayesian sparse spectrum GP (sVBSSGP) models addresses their shortcomings by adopting a Bayesian treatment of the spectral frequencies to avoid overfitting, modeling these frequencies jointly in its variational distribution to enable their interaction a posteriori, and exploiting local data for boosting the predictive performance. However, such structural improvements result in a variational lower bound that is intractable to be optimized. To resolve this, we exploit a variational parameterization trick to make it amenable to stochastic optimization. Interestingly, the resulting stochastic gradient has a linearly decomposable structure that can be exploited to refine our stochastic optimization method to incur constant time per iteration while preserving its property of being an unbiased estimator of the exact gradient of the variational lower bound. Empirical evaluation on real-world datasets shows that sVBSSGP outperforms stateof-the-art stochastic implementations of sparse GP models.

IJCAI Conference 2016 Conference Paper

DrMAD: Distilling Reverse-Mode Automatic Differentiation for Optimizing Hyperparameters of Deep Neural Networks

• Jie Fu
• Hongyin Luo
• Jiashi Feng
• Kian Hsiang Low
• Tat-Seng Chua

The performance of deep neural networks is well-known to be sensitive to the setting of their hyperparameters. Recent advances in reverse-mode automatic differentiation allow for optimizing hyperparameters with gradients. The standard way of computing these gradients involves a forward and backward pass of computations. However, the backward pass usually needs to consume unaffordable memory to store all the intermediate variables to exactly reverse the forward training procedure. In this work we propose a simple but effective method, DrMAD, to distill the knowledge of the forward pass into a shortcut path, through which we approximately reverse the training trajectory. Experiments on several image benchmark datasets show that DrMAD is at least 45 times faster and consumes 100 times less memory compared to state-of-the-art methods for optimizing hyperparameters with minimal compromise to its effectiveness. To the best of our knowledge, DrMAD is the first research attempt to make it practical to automatically tune thousands of hyperparameters of deep neural networks. The code can be downloaded from https: //github. com/bigaidream-projects/drmad

AAAI Conference 2016 Conference Paper

Gaussian Process Planning with Lipschitz Continuous Reward Functions: Towards Unifying Bayesian Optimization, Active Learning, and Beyond

• Chun Kai Ling
• Kian Hsiang Low
• Patrick Jaillet

This paper presents a novel nonmyopic adaptive Gaussian process planning (GPP) framework endowed with a general class of Lipschitz continuous reward functions that can unify some active learning/sensing and Bayesian optimization criteria and offer practitioners some flexibility to specify their desired choices for defining new tasks/problems. In particular, it utilizes a principled Bayesian sequential decision problem framework for jointly and naturally optimizing the exploration-exploitation trade-off. In general, the resulting induced GPP policy cannot be derived exactly due to an uncountable set of candidate observations. A key contribution of our work here thus lies in exploiting the Lipschitz continuity of the reward functions to solve for a nonmyopic adaptive -optimal GPP ( -GPP) policy. To plan in real time, we further propose an asymptotically optimal, branch-and-bound anytime variant of -GPP with performance guarantee. We empirically demonstrate the effectiveness of our -GPP policy and its anytime variant in Bayesian optimization and an energy harvesting task.

AAMAS Conference 2016 Conference Paper

Multi-Agent Continuous Transportation with Online Balanced Partitioning (Extended Abstract)

• Chao Wang
• Somchaya Liemhetcharat
• Kian Hsiang Low

We introduce the concept of continuous transportation task to the context of multi-agent systems. A continuous transportation task is one in which a multi-agent team visits a number of fixed locations, picks up objects, and delivers them to a transportation hub. The goal is to maximize the rate of transportation while the objects are replenished over time. In this extended abstract, we present a hybrid of centralized and distributed approaches that minimize communications in the multi-agent team. We contribute a novel online partitioning-transportation algorithm with information gathering in the multi-agent team.

AAAI Conference 2016 Conference Paper

Near-Optimal Active Learning of Multi-Output Gaussian Processes

• Yehong Zhang
• Trong Nghia Hoang
• Kian Hsiang Low
• Mohan Kankanhalli

This paper addresses the problem of active learning of a multi-output Gaussian process (MOGP) model representing multiple types of coexisting correlated environmental phenomena. In contrast to existing works, our active learning problem involves selecting not just the most informative sampling locations to be observed but also the types of measurements at each selected location for minimizing the predictive uncertainty (i. e. , posterior joint entropy) of a target phenomenon of interest given a sampling budget. Unfortunately, such an entropy criterion scales poorly in the numbers of candidate sampling locations and selected observations when optimized. To resolve this issue, we first exploit a structure common to sparse MOGP models for deriving a novel active learning criterion. Then, we exploit a relaxed form of submodularity property of our new criterion for devising a polynomial-time approximation algorithm that guarantees a constant-factor approximation of that achieved by the optimal set of selected observations. Empirical evaluation on real-world datasets shows that our proposed approach outperforms existing algorithms for active learning of MOGP and single-output GP models.

AAAI Conference 2015 Conference Paper

Parallel Gaussian Process Regression for Big Data: Low-Rank Representation Meets Markov Approximation

• Kian Hsiang Low
• Jiangbo Yu
• Jie Chen
• Patrick Jaillet

The expressive power of a Gaussian process (GP) model comes at a cost of poor scalability in the data size. To improve its scalability, this paper presents a low-rankcum-Markov approximation (LMA) of the GP model that is novel in leveraging the dual computational advantages stemming from complementing a low-rank approximate representation of the full-rank GP based on a support set of inputs with a Markov approximation of the resulting residual process; the latter approximation is guaranteed to be closest in the Kullback-Leibler distance criterion subject to some constraint and is considerably more refined than that of existing sparse GP models utilizing low-rank representations due to its more relaxed conditional independence assumption (especially with larger data). As a result, our LMA method can trade off between the size of the support set and the order of the Markov property to (a) incur lower computational cost than such sparse GP models while achieving predictive performance comparable to them and (b) accurately represent features/patterns of any scale. Interestingly, varying the Markov order produces a spectrum of LMAs with PIC approximation and full-rank GP at the two extremes. An advantage of our LMA method is that it is amenable to parallelization on multiple machines/cores, thereby gaining greater scalability. Empirical evaluation on three real-world datasets in clusters of up to 32 computing nodes shows that our centralized and parallel LMA methods are significantly more time-efficient and scalable than state-of-the-art sparse and full-rank GP regression methods while achieving comparable predictive performances.

AAAI Conference 2014 Conference Paper

GP-Localize: Persistent Mobile Robot Localization Using Online Sparse Gaussian Process Observation Model

• Nuo Xu
• Kian Hsiang Low
• Jie Chen
• Keng Kiat Lim
• Etkin Ozgul

Central to robot exploration and mapping is the task of persistent localization in environmental fields characterized by spatially correlated measurements. This paper presents a Gaussian process localization (GP-Localize) algorithm that, in contrast to existing works, can exploit the spatially correlated field measurements taken during a robot’s exploration (instead of relying on prior training data) for efficiently and scalably learning the GP observation model online through our proposed novel online sparse GP. As a result, GP-Localize is capable of achieving constant time and memory (i. e. , independent of the size of the data) per filtering step, which demonstrates the practical feasibility of using GPs for persistent robot localization and autonomy. Empirical evaluation via simulated experiments with real-world datasets and a real robot experiment shows that GP-Localize outperforms existing GP localization algorithms.

IJCAI Conference 2013 Conference Paper

A General Framework for Interacting Bayes-Optimally with Self-Interested Agents Using Arbitrary Parametric Model and Model Prior

• Trong Nghia Hoang
• Kian Hsiang Low

Recent advances in Bayesian reinforcement learning (BRL) have shown that Bayes-optimality is theoretically achievable by modeling the environment’s latent dynamics using Flat-Dirichlet- Multinomial (FDM) prior. In self-interested multiagent environments, the transition dynamics are mainly controlled by the other agent’s stochastic behavior for which FDM’s independence and modeling assumptions do not hold. As a result, FDM does not allow the other agent’s behavior to be generalized across different states nor specified using prior domain knowledge. To overcome these practical limitations of FDM, we propose a generalization of BRL to integrate the general class of parametric models and model priors, thus allowing practitioners’ domain knowledge to be exploited to produce a fine-grained and compact representation of the other agent’s behavior. Empirical evaluation shows that our approach outperforms existing multi-agent reinforcement learning algorithms.

IJCAI Conference 2013 Conference Paper

Interactive POMDP Lite: Towards Practical Planning to Predict and Exploit Intentions for Interacting with Self-Interested Agents

• Trong Nghia Hoang
• Kian Hsiang Low

A key challenge in non-cooperative multi-agent systems is that of developing efficient planning algorithms for intelligent agents to interact and perform effectively among boundedly rational, selfinterested agents (e. g. , humans). The practicality of existing works addressing this challenge is being undermined due to either the restrictive assumptions of the other agents’ behavior, the failure in accounting for their rationality, or the prohibitively expensive cost of modeling and predicting their intentions. To boost the practicality of research in this field, we investigate how intention prediction can be efficiently exploited and made practical in planning, thereby leading to efficient intention-aware planning frameworks capable of predicting the intentions of other agents and acting optimally with respect to their predicted intentions. We show that the performance losses incurred by the resulting planning policies are linearly bounded by the error of intention prediction. Empirical evaluations through a series of stochastic games demonstrate that our policies can achieve better and more robust performance than the state-of-the-art algorithms.

AAMAS Conference 2012 Conference Paper

Decentralized Active Robotic Exploration and Mapping for Probabilistic Field Classification in Environmental Sensing

• Kian Hsiang Low
• Jie Chen
• John Dolan
• Steve Chien
• David Thompson

A central problem in environmental sensing and monitoring is to classify/label the hotspots in a large-scale environmental field. This paper presents a novel \emph{decentralized active robotic exploration} (DARE) strategy for probabilistic classification/labeling of hotspots in a \emph{Gaussian process} (GP)-based field. In contrast to existing state-of-the-art exploration strategies for learning environmental field maps, the time needed to solve the DARE strategy is independent of the map resolution and the number of robots, thus making it practical for in situ, real-time active sampling. Its exploration behavior exhibits an interesting formal trade-off between that of boundary tracking until the hotspot region boundary can be accurately predicted and wide-area coverage to find new boundaries in sparsely sampled areas to be tracked. We provide a theoretical guarantee on the active exploration performance of the DARE strategy: under reasonable conditional independence assumption, we prove that it can optimally achieve two formal cost-minimizing exploration objectives based on the misclassification and entropy criteria. Importantly, this result implies that the uncertainty of labeling the hotspots in a GP-based field is greatest at or close to the hotspot region boundaries. Empirical evaluation on real-world plankton density and temperature field data shows that, subject to limited observations, DARE strategy can achieve more superior classification of hotspots and time efficiency than state-of-the-art active exploration strategies.

AAMAS Conference 2012 Conference Paper

Decision-Theoretic Approach to Maximizing Observation of Multiple Targets in Multi-Camera Surveillance

• Prabhu Natarajan
• Trong Nghia Hoang
• Kian Hsiang Low
• Mohan Kankanhalli

This paper presents a novel decision-theoretic approach to control and coordinate multiple active cameras for observing a number of moving targets in a surveillance system. This approach offers the advantages of being able to (a) account for the stochasticity of targets' motion via probabilistic modeling, and (b) address the trade-off between maximizing the expected number of observed targets and the resolution of the observed targets through stochastic optimization. One of the key issues faced by existing approaches in multi-camera surveillance is that of scalability with increasing number of targets. We show how its scalability can be improved by exploiting the problem structure: as proven analytically, our decision-theoretic approach incurs time that is linear in the number of targets to be observed during surveillance. As demonstrated empirically through simulations, our proposed approach can achieve high-quality surveillance of up to 50 targets in real time and its surveillance performance degrades gracefully with increasing number of targets. We also demonstrate our proposed approach with real AXIS 214 PTZ cameras in maximizing the number of Lego robots observed at high resolution over a surveyed rectangular area. The results are promising and clearly show the feasibility of our decision-theoretic approach in controlling and coordinating the active cameras in real surveillance system.

AAMAS Conference 2012 Conference Paper

Intention-Aware Planning under Uncertainty for Interacting with Self-Interested, Boundedly Rational Agents

• Trong Nghia Hoang
• Kian Hsiang Low

A key challenge in non-cooperative multi-agent systems is that of developing efficient planning algorithms for intelligent agents to perform effectively among boundedly rational1, self-interested (i. e. , non-cooperative) agents (e. g. , humans). To address this challenge, we investigate how intention prediction can be efficiently exploited and made practical in planning, thereby leading to efficient intention-aware planning frameworks capable of predicting the intentions of other agents and acting optimally with respect to their predicted intentions.

AAMAS Conference 2011 Conference Paper

Active Markov Information-Theoretic Path Planning for Robotic Environmental Sensing

• Kian Hsiang Low
• John M. Dolan
• Pradeep Khosla

Recent research in multi-robot exploration and mapping has focused on sampling environmental fields, which are typically modeled using the Gaussian process (GP). Existing information-theoretic exploration strategies for learning GP-based environmental field maps adopt the non-Markovian problem structure and consequently scale poorly with the length of history of observations. Hence, it becomes computationally impractical to use these strategies for in situ, real-time active sampling. To ease this computational burden, this paper presents a Markov-based approach to efficient information-theoretic path planning for active sampling of GP-based fields. We analyze the time complexity of solving the Markov-based path planning problem, and demonstrate analytically that it scales better than that of deriving the non-Markovian strategies with increasing length of planning horizon. For a class of exploration tasks called the transect sampling task, we provide theoretical guarantees on the active sampling performance of our Markov-based policy, from which ideal environmental field conditions and sampling task settings can be established to limit its performance degradation due to violation of the Markov assumption. Empirical evaluation on real-world temperature and plankton density field data shows that our Markov-based policy can generally achieve active sampling performance comparable to that of the widely-used non-Markovian greedy policies under less favorable realistic field conditions and task settings while enjoying significant computational gain over them.

AAMAS Conference 2008 Conference Paper

Adaptive Multi-Robot Wide-Area Exploration and Mapping

• Kian Hsiang Low
• John Dolan
• Pradeep Khosla

The exploration problem is a central issue in mobile robotics. A complete terrain coverage is not practical if the environment is large with only a few small hotspots. This paper presents an adaptive multi-robot exploration strategy that is novel in performing both wide-area coverage and hotspot sampling using non-myopic path planning. As a result, the environmental phenomena can be accurately mapped. It is based on a dynamic programming formulation, which we call the Multi-robot Adaptive Sampling Problem (MASP). A key feature of MASP is in covering the entire adaptivity spectrum, thus allowing strategies of varying adaptivity to be formed and theoretically analyzed in their performance; a more adaptive strategy improves mapping accuracy. We apply MASP to sampling the Gaussian and log- Gaussian processes, and analyze if the resulting strategies are adaptive and maximize wide-area coverage and hotspot sampling. Solving MASP is non-trivial as it comprises continuous state components. So, it is reformulated for convex analysis, which allows discrete-state monotone-bounding approximation to be developed. We provide a theoretical guarantee on the policy quality of the approximate MASP (aMASP) for using in MASP. Although aMASP can be solved exactly, its state size grows exponentially with the number of stages. To alleviate this computational difficulty, anytime algorithms are proposed based on aMASP, one of which can guarantee its policy quality for MASP in real time.

AAAI Conference 2004 Conference Paper

Task Allocation via Self-Organizing Swarm Coalitions in Distributed Mobile Sensor Network

• Kian Hsiang Low
• and Marcelo H. Ang

This paper presents a task allocation scheme via selforganizing swarm coalitions for distributed mobile sensor network coverage. Our approach uses the concepts of ant behavior to self-regulate the regional distributions of sensors in proportion to that of the moving targets to be tracked in a non-stationary environment. As a result, the adverse effects of task interference between robots are minimized and sensor network coverage is improved. Quantitative comparisons with other tracking strategies such as static sensor placement, potential fields, and auction-based negotiation show that our approach can provide better coverage and greater flexibility to respond to environmental changes.

IJCAI Conference 2003 Conference Paper

Action Selection for Single- and Multi-Robot Tasks Using Cooperative Extended Kohonen Maps

• Kian Hsiang Low
• Wee Kheng Leow
• Marcelo H. Ang
• Jr

This paper presents an action selection framework based on an assemblage of self-organizing neural networks called Cooperative Extended Kohonen Maps. This framework encapsulates two features that significantly enhance a robot's action selection capability: self-organization in the continuous state and action spaces to provide smooth, efficient and fine motion control; action selection via the cooperation and competition of Extended Kohonen Maps to achieve more complex motion tasks. Qualitative and quantitative comparisons for single- and multirobot tasks show our framework can provide better action selection than do potential fields method.

AAMAS Conference 2003 Conference Paper

Action Selection in Continuous State and Action Spaces by Cooperation and Competition of Extended Kohonen Maps

• Kian Hsiang Low
• Wee Kheng Leow
• Marcelo H. Ang Jr.

AAMAS Conference 2002 Conference Paper

A Hybrid Mobile Robot Architecture with Integrated Planning and Control

• Kian Hsiang Low
• Wee Kheng Leow
• Marcelo H. Ang
• Jr.