arXiv subjects

Kihwan Kim

Publications and source records attributed to Kihwan Kim.

At least 37 records · Page 2Linked to original sources

arXiv2021

An Empirical Analysis on Transparent Algorithmic Exploration in Recommender Systems

All learning algorithms for recommendations face inevitable and critical trade-off between exploiting partial knowledge of a user's preferences for short-term satisfaction and exploring additional user preferences for long-term coverage. Although exploration is indispensable for long success of a recommender system, the exploration has been considered as the risk to decrease user satisfaction. The reason for the risk is that items chosen for exploration frequently mismatch with the user's interests. To mitigate this risk, recommender systems have mixed items chosen for exploration into a recommendation list, disguising the items as recommendations to elicit feedback on the items to discover the user's additional tastes. This mix-in approach has been widely used in many recommenders, but there is rare research, evaluating the effectiveness of the mix-in approach or proposing a new approach for eliciting user feedback without deceiving users. In this work, we aim to propose a new approach for feedback elicitation without any deception and compare our approach to the conventional mix-in approach for evaluation. To this end, we designed a recommender interface that reveals which items are for exploration and conducted a within-subject study with 94 MTurk workers. Our results indicated that users left significantly more feedback on items chosen for exploration with our interface. Besides, users evaluated that our new interface is better than the conventional mix-in interface in terms of novelty, diversity, transparency, trust, and satisfaction. Finally, path analysis show that, in only our new interface, exploration caused to increase user-centric evaluation metrics. Our work paves the way for how to design an interface, which utilizes learning algorithm based on users' feedback signals, giving better user experience and gathering more feedback data.

cs.IR↗

arXiv2021

Double-EIT Ground-State Cooling of Stationary Two-Dimensional Ion Lattices

We theoretically and experimentally investigate double electromagnetically induced transparency (double-EIT) cooling of two-dimensional ion crystals confined in a Paul trap. The double-EIT ground-state cooling is observed for \Yb ions with clock state, for which EIT cooling has not been realized like many other ions with a simple $Λ$-scheme. A cooling rate of $\dot{\bar n}=34~(\pm1.8)~\rm{ms}^{-1}$ and a cooling limit of $\bar n=0.06~(\pm 0.059)$ are observed for a single ion. The measured cooling rate and limit are consistent with theoretical predictions. We apply double-EIT cooling to the transverse modes of two-dimensional (2D) crystals with up to 12 ions. In our 2D crystals, the micromotion and the transverse mode directions are perpendicular, which makes them decoupled. Therefore, the cooling on transverse modes is not disturbed by micromotion, which is confirmed in our experiment. For the center of mass mode of a 12 ions crystal, we observe a cooling rate and a cooling limit that are consistent with those of a single ion, including heating rates proportional to the number of ions. This method can be extended to other hyperfine qubits, and near ground-state cooling of stationary 2D crystals with large numbers of ions may advance the field of quantum information sciences.

quant-ph↗

arXiv2021

Quantum computation and simulation with vibrational modes of trapped ions

Vibrational degrees of freedom in trapped-ion systems have recently been gaining attention as a quantum resource, beyond the role as a mediator for entangling quantum operations on internal degrees of freedom, because of the large available Hilbert space. The vibrational modes can be represented as quantum harmonic oscillators and thus offer a Hilbert space with infinite dimension. Here we review recent theoretical and experimental progress in the coherent manipulation of the vibrational modes, including bosonic encoding schemes in quantum information, reliable and efficient measurement techniques, and quantum operations that allow various quantum simulations and quantum computation algorithms. We describe experiments using the vibrational modes, including the preparation of non-classical states, molecular vibronic sampling, and applications in quantum thermodynamics. We finally discuss the potential prospects and challenges of trapped-ion vibrational-mode quantum information processing.

quant-ph↗

arXiv2021

Neural 3D Clothes Retargeting from a Single Image

In this paper, we present a method of clothes retargeting; generating the potential poses and deformations of a given 3D clothing template model to fit onto a person in a single RGB image. The problem is fundamentally ill-posed as attaining the ground truth data is impossible, i.e., images of people wearing the different 3D clothing template model at exact same pose. We address this challenge by utilizing large-scale synthetic data generated from physical simulation, allowing us to map 2D dense body pose to 3D clothing deformation. With the simulated data, we propose a semi-supervised learning framework that validates the physical plausibility of the 3D deformation by matching with the prescribed body-to-cloth contact points and clothing silhouette to fit onto the unlabeled real images. A new neural clothes retargeting network (CRNet) is designed to integrate the semi-supervised retargeting task in an end-to-end fashion. In our evaluation, we show that our method can predict the realistic 3D pose and deformation field needed for retargeting clothes models in real-world examples.

cs.CV↗

arXiv2020

Online Adaptation for Consistent Mesh Reconstruction in the Wild

This paper presents an algorithm to reconstruct temporally consistent 3D meshes of deformable object instances from videos in the wild. Without requiring annotations of 3D mesh, 2D keypoints, or camera pose for each video frame, we pose video-based reconstruction as a self-supervised online adaptation problem applied to any incoming test video. We first learn a category-specific 3D reconstruction model from a collection of single-view images of the same category that jointly predicts the shape, texture, and camera pose of an image. Then, at inference time, we adapt the model to a test video over time using self-supervised regularization terms that exploit temporal consistency of an object instance to enforce that all reconstructed meshes share a common texture map, a base shape, as well as parts. We demonstrate that our algorithm recovers temporally consistent and reliable 3D structures from videos of non-rigid objects including those of animals captured in the wild -- an extremely challenging task rarely addressed before.

cs.CV↗

arXiv2020

ST-GRAT: A Novel Spatio-temporal Graph Attention Network for Accurately Forecasting Dynamically Changing Road Speed

Predicting road traffic speed is a challenging task due to different types of roads, abrupt speed change and spatial dependencies between roads; it requires the modeling of dynamically changing spatial dependencies among roads and temporal patterns over long input sequences. This paper proposes a novel spatio-temporal graph attention (ST-GRAT) that effectively captures the spatio-temporal dynamics in road networks. The novel aspects of our approach mainly include spatial attention, temporal attention, and spatial sentinel vectors. The spatial attention takes the graph structure information (e.g., distance between roads) and dynamically adjusts spatial correlation based on road states. The temporal attention is responsible for capturing traffic speed changes, and the sentinel vectors allow the model to retrieve new features from spatially correlated nodes or preserve existing features. The experimental results show that ST-GRAT outperforms existing models, especially in difficult conditions where traffic speeds rapidly change (e.g., rush hours). We additionally provide a qualitative study to analyze when and where ST-GRAT tended to make accurate predictions during rush-hour times.

cs.LG↗

arXiv2020

DeepGMR: Learning Latent Gaussian Mixture Models for Registration

Point cloud registration is a fundamental problem in 3D computer vision, graphics and robotics. For the last few decades, existing registration algorithms have struggled in situations with large transformations, noise, and time constraints. In this paper, we introduce Deep Gaussian Mixture Registration (DeepGMR), the first learning-based registration method that explicitly leverages a probabilistic registration paradigm by formulating registration as the minimization of KL-divergence between two probability distributions modeled as mixtures of Gaussians. We design a neural network that extracts pose-invariant correspondences between raw point clouds and Gaussian Mixture Model (GMM) parameters and two differentiable compute blocks that recover the optimal transformation from matched GMM parameters. This construction allows the network learn an SE(3)-invariant feature space, producing a global registration method that is real-time, generalizable, and robust to noise. Across synthetic and real-world data, our proposed method shows favorable performance when compared with state-of-the-art geometry-based and learning-based registration methods.

cs.CV↗

arXiv2020

Single ion-qubit exceeding one hour coherence time

Realizing a long coherence time quantum memory is a major challenge of current quantum technology. Here, we report a single \Yb ion-qubit memory with over one hour coherence time, an order of improvement compared to the state-of-the-art record. The long coherence time memory is realized by addressing various technical challenges such as ambient magnetic-field noise, phase noise and leakage of the microwave oscillator. Moreover, systematically study the decoherence process of our quantum memory by quantum process tomography, which enables to apply the strict criteria of quantum coherence, relative entropy of coherence. We also benchmark our quantum memory by its ability in preserving quantum information, i.e., the robustness of quantum memory, which clearly shows that over 6000 s, our quantum memory preserves non-classical quantum information. Our results verify the stability of the quantum memory in hours level and indicate its versatile applicability in various scenarios.

quant-ph↗

arXiv2020

Bi3D: Stereo Depth Estimation via Binary Classifications

Stereo-based depth estimation is a cornerstone of computer vision, with state-of-the-art methods delivering accurate results in real time. For several applications such as autonomous navigation, however, it may be useful to trade accuracy for lower latency. We present Bi3D, a method that estimates depth via a series of binary classifications. Rather than testing if objects are at a particular depth $D$, as existing stereo methods do, it classifies them as being closer or farther than $D$. This property offers a powerful mechanism to balance accuracy and latency. Given a strict time budget, Bi3D can detect objects closer than a given distance in as little as a few milliseconds, or estimate depth with arbitrarily coarse quantization, with complexity linear with the number of quantization levels. Bi3D can also use the allotted quantization levels to get continuous depth, but in a specific depth range. For standard stereo (i.e., continuous depth on the whole range), our method is close to or on par with state-of-the-art, finely tuned stereo methods.

cs.CV↗

arXiv2020

Novel View Synthesis of Dynamic Scenes with Globally Coherent Depths from a Monocular Camera

This paper presents a new method to synthesize an image from arbitrary views and times given a collection of images of a dynamic scene. A key challenge for the novel view synthesis arises from dynamic scene reconstruction where epipolar geometry does not apply to the local motion of dynamic contents. To address this challenge, we propose to combine the depth from single view (DSV) and the depth from multi-view stereo (DMV), where DSV is complete, i.e., a depth is assigned to every pixel, yet view-variant in its scale, while DMV is view-invariant yet incomplete. Our insight is that although its scale and quality are inconsistent with other views, the depth estimation from a single view can be used to reason about the globally coherent geometry of dynamic contents. We cast this problem as learning to correct the scale of DSV, and to refine each depth with locally consistent motions between views to form a coherent depth estimation. We integrate these tasks into a depth fusion network in a self-supervised fashion. Given the fused depth maps, we synthesize a photorealistic virtual view in a specific location and time with our deep blending network that completes the scene and renders the virtual view. We evaluate our method of depth estimation and view synthesis on diverse real-world dynamic scenes and show the outstanding performance over existing methods.

cs.CV↗

arXiv2020

Two-shot Spatially-varying BRDF and Shape Estimation

Capturing the shape and spatially-varying appearance (SVBRDF) of an object from images is a challenging task that has applications in both computer vision and graphics. Traditional optimization-based approaches often need a large number of images taken from multiple views in a controlled environment. Newer deep learning-based approaches require only a few input images, but the reconstruction quality is not on par with optimization techniques. We propose a novel deep learning architecture with a stage-wise estimation of shape and SVBRDF. The previous predictions guide each estimation, and a joint refinement network later refines both SVBRDF and shape. We follow a practical mobile image capture setting and use unaligned two-shot flash and no-flash images as input. Both our two-shot image capture and network inference can run on mobile hardware. We also create a large-scale synthetic training dataset with domain-randomized geometry and realistic materials. Extensive experiments on both synthetic and real-world datasets show that our network trained on a synthetic dataset can generalize well to real-world images. Comparisons with recent approaches demonstrate the superior performance of the proposed approach.

cs.CV↗

arXiv2020

Self-supervised Single-view 3D Reconstruction via Semantic Consistency

We learn a self-supervised, single-view 3D reconstruction model that predicts the 3D mesh shape, texture and camera pose of a target object with a collection of 2D images and silhouettes. The proposed method does not necessitate 3D supervision, manually annotated keypoints, multi-view images of an object or a prior 3D template. The key insight of our work is that objects can be represented as a collection of deformable parts, and each part is semantically coherent across different instances of the same category (e.g., wings on birds and wheels on cars). Therefore, by leveraging self-supervisedly learned part segmentation of a large collection of category-specific images, we can effectively enforce semantic consistency between the reconstructed meshes and the original images. This significantly reduces ambiguities during joint prediction of shape and camera pose of an object, along with texture. To the best of our knowledge, we are the first to try and solve the single-view reconstruction problem without a category-specific template mesh or semantic keypoints. Thus our model can easily generalize to various object categories without such labels, e.g., horses, penguins, etc. Through a variety of experiments on several categories of deformable and rigid objects, we demonstrate that our unsupervised method performs comparably if not better than existing category-specific reconstruction methods learned with supervision.

cs.CV↗

arXiv2019

Realization of two-dimensional crystal of ions in a monolithic Paul trap

We present a simple Paul trap that stably accommodates up to a couple of dozens of \ensuremath{^{171}\mathrm{Yb}^+~} ions in a stationary two-dimensional lattice. The trap is constructed on a single plate of gold-plated laser-machined alumina and can produce a pancake-like pseudo-potential that makes ions form a self-assembly two-dimensional crystal which locates on the plane composed of axial and one of the transverse axes with around 5 $μ$m spacing. We use Raman laser beams to coherently manipulate these ion-qubits where the net propagation direction is perpendicular to the plane of the crystal and micromotion. We perform the coherent operations and study the spectrum of vibrational modes through globally addressed Raman laser-beams on a dozen of ions in the two-dimensional crystal. We measure the amplitude of micro-motion by comparing the strengths of carrier and micro-motion sideband transitions with three ions, where the micro-motion amplitude is similar to that of a single ion. The spacings of ions are small enough for large coupling strengths, which is a favorable condition for two-dimensional quantum simulation.

quant-ph↗

arXiv2019

Neural Inverse Rendering of an Indoor Scene from a Single Image

Inverse rendering aims to estimate physical attributes of a scene, e.g., reflectance, geometry, and lighting, from image(s). Inverse rendering has been studied primarily for single objects or with methods that solve for only one of the scene attributes. We propose the first learning-based approach that jointly estimates albedo, normals, and lighting of an indoor scene from a single image. Our key contribution is the Residual Appearance Renderer (RAR), which can be trained to synthesize complex appearance effects (e.g., inter-reflection, cast shadows, near-field illumination, and realistic shading), which would be neglected otherwise. This enables us to perform self-supervised learning on real data using a reconstruction loss, based on re-synthesizing the input image from the estimated components. We finetune with real data after pretraining with synthetic data. To this end, we use physically-based rendering to create a large-scale synthetic dataset, which is a significant improvement over prior datasets. Experimental results show that our approach outperforms state-of-the-art methods that estimate one or more scene attributes.

cs.CV↗

arXiv2019

Modular Quantum Computation in a Trapped Ion System

Modern computation relies crucially on modular architectures, breaking a complex algorithm into self-contained subroutines. A client can then call upon a remote server to implement parts of the computation independently via an application programming interface (API). Present APIs relay only classical information. Here we implement a quantum API that enables a client to estimate the absolute value of the trace of a server-provided unitary $U$. We demonstrate that the algorithm functions correctly irrespective of what unitary $U$ the server implements or how the server specifically realizes $U$. Our experiment involves pioneering techniques to coherently swap qubits encoded within the motional states of a trapped \Yb ion, controlled on its hyperfine state. This constitutes the first demonstration of modular computation in the quantum regime, providing a step towards scalable, parallelization of quantum computation.

quant-ph↗

arXiv2019

Shortcuts to adiabaticity by counterdiabatic driving for trapped-ion displacement in phase space

The application of adiabatic protocols in quantum technologies is severely limited by environmental sources of noise and decoherence. Shortcuts to adiabaticity by counterdiabatic driving constitute a powerful alternative that speed up time-evolution while mimicking adiabatic dynamics. Here we present the first experimental implementation of counterdiabatic driving in a continuous variable system, a shortcut to the adiabatic transport of a trapped ion in the phase space. The resulting dynamics is equivalent to a "fast-motion video" of the adiabatic trajectory. The robustness of this protocol is shown to surpass that of competing schemes based on classical local controls and Fourier optimization methods. Our results demonstrate that shortcuts to adiabaticity provide a robust speedup of quantum protocols of wide applicability in quantum technologies.

quant-ph↗

arXiv2019

Error-Mitigated Quantum Gates Exceeding Physical Fidelities in a Trapped-Ion System

Various quantum applications can be reduced to estimating expectation values, which are inevitably deviated by operational and environmental errors. Although errors can be tackled by quantum error correction, the overheads are far from being affordable for near-term technologies. To alleviate the detrimental effects of errors, quantum error mitigation techniques have been proposed, which require no additional qubit resources. Here, we benchmark the performance of a quantum error mitigation technique based on probabilistic error cancellation in a trapped-ion system. Our results clearly show that effective gate fidelities exceed physical fidelities, i.e. we surpass the break-even point of eliminating gate errors, by programming quantum circuits. The error rates are effectively reduced from $(1.10\pm 0.12)\times10^{-3}$ to $(1.44\pm 5.28)\times10^{-5}$ and from $(0.99\pm 0.06)\times10^{-2}$ to $(0.96\pm 0.10)\times10^{-3}$ for single- and two-qubit gates, respectively. Our demonstration opens up the possibility of implementing high-fidelity computations on a near-term noisy quantum device.

quant-ph↗

arXiv2019

Putting Humans in a Scene: Learning Affordance in 3D Indoor Environments

Affordance modeling plays an important role in visual understanding. In this paper, we aim to predict affordances of 3D indoor scenes, specifically what human poses are afforded by a given indoor environment, such as sitting on a chair or standing on the floor. In order to predict valid affordances and learn possible 3D human poses in indoor scenes, we need to understand the semantic and geometric structure of a scene as well as its potential interactions with a human. To learn such a model, a large-scale dataset of 3D indoor affordances is required. In this work, we build a fully automatic 3D pose synthesizer that fuses semantic knowledge from a large number of 2D poses extracted from TV shows as well as 3D geometric knowledge from voxel representations of indoor scenes. With the data created by the synthesizer, we introduce a 3D pose generative model to predict semantically plausible and physically feasible human poses within a given scene (provided as a single RGB, RGB-D, or depth image). We demonstrate that our human affordance prediction method consistently outperforms existing state-of-the-art methods.

cs.CV↗