arXiv subjects

Minkyu Kim

Publications and source records attributed to Minkyu Kim.

At least 73 records · Page 4Linked to original sources

arXiv2021

SmoothMix: Training Confidence-calibrated Smoothed Classifiers for Certified Robustness

Randomized smoothing is currently a state-of-the-art method to construct a certifiably robust classifier from neural networks against $\ell_2$-adversarial perturbations. Under the paradigm, the robustness of a classifier is aligned with the prediction confidence, i.e., the higher confidence from a smoothed classifier implies the better robustness. This motivates us to rethink the fundamental trade-off between accuracy and robustness in terms of calibrating confidences of a smoothed classifier. In this paper, we propose a simple training scheme, coined SmoothMix, to control the robustness of smoothed classifiers via self-mixup: it trains on convex combinations of samples along the direction of adversarial perturbation for each input. The proposed procedure effectively identifies over-confident, near off-class samples as a cause of limited robustness in case of smoothed classifiers, and offers an intuitive way to adaptively set a new decision boundary between these samples for better robustness. Our experimental results demonstrate that the proposed method can significantly improve the certified $\ell_2$-robustness of smoothed classifiers compared to existing state-of-the-art robust training methods.

cs.LG↗

arXiv2021

Global-Local Item Embedding for Temporal Set Prediction

Temporal set prediction is becoming increasingly important as many companies employ recommender systems in their online businesses, e.g., personalized purchase prediction of shopping baskets. While most previous techniques have focused on leveraging a user's history, the study of combining it with others' histories remains untapped potential. This paper proposes Global-Local Item Embedding (GLOIE) that learns to utilize the temporal properties of sets across whole users as well as within a user by coining the names as global and local information to distinguish the two temporal patterns. GLOIE uses Variational Autoencoder (VAE) and dynamic graph-based model to capture global and local information and then applies attention to integrate resulting item embeddings. Additionally, we propose to use Tweedie output for the decoder of VAE as it can easily model zero-inflated and long-tailed distribution, which is more suitable for several real-world data distributions than Gaussian or multinomial counterparts. When evaluated on three public benchmarks, our algorithm consistently outperforms previous state-of-the-art methods in most ranking metrics.

cs.LG↗

arXiv2021

Information-Theoretic Based Target Search with Multiple Agents

This paper proposes an online path planning and motion generation algorithm for heterogeneous robot teams performing target search in a real-world environment. Path selection for each robot is optimized using an information-theoretic formulation and is computed sequentially for each agent. First, we generate candidate trajectories sampled from both global waypoints derived from vertical cell decomposition and local frontier points. From this set, we choose the path with maximum information gain. We demonstrate that the hierarchical sequential decision-making structure provided by the algorithm is scalable to multiple agents in a simulation setup. We also validate our framework in a real-world apartment setting using a two robot team comprised of the Unitree A1 quadruped and the Toyota HSR mobile manipulator searching for a person. The agents leverage an efficient leader-follower communication structure where only critical information is shared.

cs.RO↗

arXiv2021

One4all User Representation for Recommender Systems in E-commerce

General-purpose representation learning through large-scale pre-training has shown promising results in the various machine learning fields. For an e-commerce domain, the objective of general-purpose, i.e., one for all, representations would be efficient applications for extensive downstream tasks such as user profiling, targeting, and recommendation tasks. In this paper, we systematically compare the generalizability of two learning strategies, i.e., transfer learning through the proposed model, ShopperBERT, vs. learning from scratch. ShopperBERT learns nine pretext tasks with 79.2M parameters from 0.8B user behaviors collected over two years to produce user embeddings. As a result, the MLPs that employ our embedding method outperform more complex models trained from scratch for five out of six tasks. Specifically, the pre-trained embeddings have superiority over the task-specific supervised features and the strong baselines, which learn the auxiliary dataset for the cold-start problem. We also show the computational efficiency and embedding visualization of the pre-trained features.

cs.IR↗

arXiv2021

A pair of homotopy-theoretic version of TQFT's induced by a Brown functor

The purpose of this paper is to study some obstruction classes induced by a construction of a homotopy-theoretic version of projective TQFT (projective HTQFT for short). A projective HTQFT is given by a symmetric monoidal projective functor whose domain is the cospan category of pointed finite CW-spaces instead of a cobordism category. We construct a pair of projective HTQFT's starting from a $\mathsf{Hopf}^\mathsf{bc}_k$-valued Brown functor where $\mathsf{Hopf}^\mathsf{bc}_k$ is the category of bicommutative Hopf algebras over a field $k$ : the cospanical path-integral and the spanical path-integral of the Brown functor. They induce obstruction classes by an analogue of the second cohomology class associated with projective representations. In this paper, we derive some formulae of those obstruction classes. We apply the formulae to prove that the dimension reduction of the cospanical and spanical path-integrals are lifted to HTQFT's. In another application, we reproduce the Dijkgraaf-Witten TQFT and the Turaev-Viro TQFT from an ordinary $\mathsf{Hopf}^\mathsf{bc}_k$-valued homology theory.

math.AT↗

arXiv2021

Kitaev's stabilizer code and chain complex theory of bicommutative Hopf algebras

In this paper, we give a generalization of Kitaev's stabilizer code based on chain complex theory of bicommutative Hopf algebras. Due to the bicommutativity, the Kitaev's stabilizer code extends to a broader class of spaces, e.g. finite CW-complexes ; more generally short abstract complex over a commutative unital ring R which is introduced in this paper. Given a finite-dimensional bisemisimple bicommutative Hopf algebra with an R-action, we introduce some analogues of A-stabilizers, B-stabilizers and the local Hamiltonian, which we call by the (+)-stabilizers, the (-)-stabilizers and the elementary operator respectively. We prove that the eigenspaces of the elementary operator give an orthogonal decomposition and the ground-state space is isomorphic to the homology Hopf algebra. In application to topology, we propose a formulation of topological local stabilizer models in a functorial way. It is known that the ground-state spaces of Kitaev's stabilizer code extends to Turaev-Viro TQFT. We prove that the 0-eigenspaces of a topological local stabilizer model extends to a projective TQFT which is improved to a TQFT in typical examples. Furthermore, we give a generalization of the duality in the literature based on the Poincar'e-Lefschetz duality of R-oriented manifolds.

math-ph↗

arXiv2020

Integrals along bimonoid homomorphisms

We introduce a notion of an integral along a bimonoid homomorphism as a simultaneous generalization of the integral and cointegral of bimonoids. The purpose of this paper is to characterize an existence of a specific integral, called a normalized generator integral, along a bimonoid homomorphism in terms of the kernel and cokernel of the homomorphism. We introduce a notion of a volume on an abelian category as a generalization of the dimension of vector spaces and the order of abelian groups. In applications, we show that there exists a nontrivial volume partially defined on a category of bicommutative Hopf monoids. The volume yields a notion of Fredholm homomorphisms between bicommutative Hopf monoids, which gives an analogue of the Fredholm index theory. This paper gives a technical preliminary of our subsequent paper about a construction of TQFTs.

math.QA↗

arXiv2020

Homology theory valued in the category of bicommutative Hopf algebras

The codomain category of a generalized homology theory is the category of modules over a ring. For an abelian category A, an A-valued (generalized) homology theory is defined by formally replacing the category of modules with the category A. It is known that the category of bicommutative (i.e. commutative and cocommutative) Hopf algebras over a field k is an abelian category. Denote the category by H. In this paper, we give some ways to construct H-valued homology theories. As a main result, we give H-valued homology theories whose coefficients are neither group Hopf algebras nor function Hopf algebras. The examples contain not only ordinary homology theories but also extraordinary ones.

math.AT↗

arXiv2019

On-demand generation and characterization of a microwave time-bin qubit

Superconducting circuits offer a scalable platform for the construction of large-scale quantum networks where information can be encoded in multiple temporal modes of propagating microwaves. Characterization of such microwave signals with a method extendable to an arbitrary number of temporal modes with a single detector and demonstration of their phase-robust nature are of great interest. Here we show the on-demand generation and Wigner tomography of a microwave time-bin qubit with superconducting circuit quantum electrodynamics architecture. We perform the tomography with a single heterodyne detector by dynamically changing the measurement quadrature with a phase-sensitive amplifier independently for the two temporal modes. By generating and measuring the qubits with hardware lacking a shared phase reference, we demonstrate conservation of phase information in each time-bin qubit generated.

quant-ph↗

arXiv2019

Lick Observatory Supernova Search Follow-Up Program: Photometry Data Release of 93 Type Ia Supernovae

We present BVRI and unfiltered light curves of 93 Type Ia supernovae (SNe Ia) from the Lick Observatory Supernova Search (LOSS) follow-up program conducted between 2005 and 2018. Our sample consists of 78 spectroscopically normal SNe Ia, with the remainder divided between distinct subclasses (three SN 1991bg-like, three SN 1991T-like, four SNe Iax, two peculiar, and three super-Chandrasekhar events), and has a median redshift of 0.0192. The SNe in our sample have a median coverage of 16 photometric epochs at a cadence of 5.4 days, and the median first observed epoch is ~4.6 days before maximum B-band light. We describe how the SNe in our sample are discovered, observed, and processed, and we compare the results from our newly developed automated photometry pipeline to those from the previous processing pipeline used by LOSS. After investigating potential biases, we derive a final systematic uncertainty of 0.03 mag in BVRI for our dataset. We perform an analysis of our light curves with particular focus on using template fitting to measure the parameters that are useful in standardising SNe Ia as distance indicators. All of the data are available to the community, and we encourage future studies to incorporate our light curves in their analyses.

astro-ph.SR↗

arXiv2019

Tripartite Heterogeneous Graph Propagation for Large-scale Social Recommendation

Graph Neural Networks (GNNs) have been emerging as a promising method for relational representation including recommender systems. However, various challenging issues of social graphs hinder the practical usage of GNNs for social recommendation, such as their complex noisy connections and high heterogeneity. The oversmoothing of GNNs is an obstacle of GNN-based social recommendation as well. Here we propose a new graph embedding method Heterogeneous Graph Propagation (HGP) to tackle these issues. HGP uses a group-user-item tripartite graph as input to reduce the number of edges and the complexity of paths in a social graph. To solve the oversmoothing issue, HGP embeds nodes under a personalized PageRank based propagation scheme, separately for group-user graph and user-item graph. Node embeddings from each graph are integrated using an attention mechanism. We evaluate our HGP on a large-scale real-world dataset consisting of 1,645,279 nodes and 4,711,208 edges. The experimental results show that HGP outperforms several baselines in terms of AUC and F1-score metrics.

cs.SI↗

arXiv2019

Toward Achieving Formal Guarantees for Human-Aware Controllers in Human-Robot Interactions

With the primary objective of human-robot interaction being to support humans' goals, there exists a need to formally synthesize robot controllers that can provide the desired service. Synthesis techniques have the benefit of providing formal guarantees for specification satisfaction. There is potential to apply these techniques for devising robot controllers whose specifications are coupled with human needs. This paper explores the use of formal methods to construct human-aware robot controllers to support the productivity requirements of humans. We tackle these types of scenarios via human workload-informed models and reactive synthesis. This strategy allows us to synthesize controllers that fulfill formal specifications that are expressed as linear temporal logic formulas. We present a case study in which we reason about a work delivery and pickup task such that the robot increases worker productivity, but not stress induced by high work backlog. We demonstrate our controller using the Toyota HSR, a mobile manipulator robot. The results demonstrate the realization of a robust robot controller that is guaranteed to properly reason and react in collaborative tasks with human partners.

cs.RO↗

arXiv2019

A generalization of Dijkgraaf-Witten theory

The main purpose of this paper is to give a generalization of Dijkgraaf-Witten theory. Consider a morphism from a smash product of spectra E,F to another spectrum G. We construct a TQFT for E-oriented manifolds using a representative of an F-cohomology class of the classifying space of a finite group. For the case that each of spectra E,F,G is given as the K-theory spectrum KU, we further generalize our construction based on non-commutative settings.

math.AT↗

arXiv2018

LAAIR: A Layered Architecture for Autonomous Interactive Robots

When developing general purpose robots, the overarching software architecture can greatly affect the ease of accomplishing various tasks. Initial efforts to create unified robot systems in the 1990s led to hybrid architectures, emphasizing a hierarchy in which deliberative plans direct the use of reactive skills. However, since that time there has been significant progress in the low-level skills available to robots, including manipulation and perception, making it newly feasible to accomplish many more tasks in real-world domains. There is thus renewed optimism that robots will be able to perform a wide array of tasks while maintaining responsiveness to human operators. However, the top layer in traditional hybrid architectures, designed to achieve long-term goals, can make it difficult to react quickly to human interactions during goal-driven execution. To mitigate this difficulty, we propose a novel architecture that supports such transitions by adding a top-level reactive module which has flexible access to both reactive skills and a deliberative control module. To validate this architecture, we present a case study of its application on a domestic service robot platform.

cs.RO↗

arXiv2018

CHOPT : Automated Hyperparameter Optimization Framework for Cloud-Based Machine Learning Platforms

Many hyperparameter optimization (HyperOpt) methods assume restricted computing resources and mainly focus on enhancing performance. Here we propose a novel cloud-based HyperOpt (CHOPT) framework which can efficiently utilize shared computing resources while supporting various HyperOpt algorithms. We incorporate convenient web-based user interfaces, visualization, and analysis tools, enabling users to easily control optimization procedures and build up valuable insights with an iterative analysis procedure. Furthermore, our framework can be incorporated with any cloud platform, thus complementarily increasing the efficiency of conventional deep learning frameworks. We demonstrate applications of CHOPT with tasks such as image recognition and question-answering, showing that our framework can find hyperparameter configurations competitive with previous work. We also show CHOPT is capable of providing interesting observations through its analysing tools

cs.LG↗

arXiv2018

NSML: Meet the MLaaS platform with a real-world case study

The boom of deep learning induced many industries and academies to introduce machine learning based approaches into their concern, competitively. However, existing machine learning frameworks are limited to sufficiently fulfill the collaboration and management for both data and models. We proposed NSML, a machine learning as a service (MLaaS) platform, to meet these demands. NSML helps machine learning work be easily launched on a NSML cluster and provides a collaborative environment which can afford development at enterprise scale. Finally, NSML users can deploy their own commercial services with NSML cluster. In addition, NSML furnishes convenient visualization tools which assist the users in analyzing their work. To verify the usefulness and accessibility of NSML, we performed some experiments with common examples. Furthermore, we examined the collaborative advantages of NSML through three competitions with real-world use cases.

cs.DC↗

arXiv2018

Social Navigation Planning Based on People's Awareness of Robots

When mobile robots maneuver near people, they run the risk of rudely blocking their paths; but not all people behave the same around robots. People that have not noticed the robot are the most difficult to predict. This paper investigates how mobile robots can generate acceptable paths in dynamic environments by predicting human behavior. Here, human behavior may include both physical and mental behavior, we focus on the latter. We introduce a simple safe interaction model: when a human seems unaware of the robot, it should avoid going too close. In this study, people around robots are detected and tracked using sensor fusion and filtering techniques. To handle uncertainties in the dynamic environment, a Partially-Observable Markov Decision Process Model (POMDP) is used to formulate a navigation planning problem in the shared environment. People's awareness of robots is inferred and included as a state and reward model in the POMDP. The proposed planner enables a robot to change its navigation plan based on its perception of each person's robot-awareness. As far as we can tell, this is a new capability. We conduct simulation and experiments using the Toyota Human Support Robot (HSR) to validate our approach. We demonstrate that the proposed framework is capable of running in real-time.

cs.RO↗

arXiv2018

An Architecture for Person-Following using Active Target Search

This paper addresses a novel architecture for person-following robots using active search. The proposed system can be applied in real-time to general mobile robots for learning features of a human, detecting and tracking, and finally navigating towards that person. To succeed at person-following, perception, planning, and robot behavior need to be integrated properly. Toward this end, an active target searching capability, including prediction and navigation toward vantage locations for finding human targets, is proposed. The proposed capability aims at improving the robustness and efficiency for tracking and following people under dynamic conditions such as crowded environments. A multi-modal sensor information approach including fusing an RGB-D sensor and a laser scanner, is pursued to robustly track and identify human targets. Bayesian filtering for keeping track of human and a regression algorithm to predict the trajectory of people are investigated. In order to make the robot autonomous, the proposed framework relies on a behavior-tree structure. Using Toyota Human Support Robot (HSR), real-time experiments demonstrate that the proposed architecture can generate fast, efficient person-following behaviors.

cs.RO↗