arXiv subjects

Minji Kim

Publications and source records attributed to Minji Kim.

33 records · Page 2Linked to original sources

arXiv2024

Physics-informed reduced order model with conditional neural fields

This study presents the conditional neural fields for reduced-order modeling (CNF-ROM) framework to approximate solutions of parametrized partial differential equations (PDEs). The approach combines a parametric neural ODE (PNODE) for modeling latent dynamics over time with a decoder that reconstructs PDE solutions from the corresponding latent states. We introduce a physics-informed learning objective for CNF-ROM, which includes two key components. First, the framework uses coordinate-based neural networks to calculate and minimize PDE residuals by computing spatial derivatives via automatic differentiation and applying the chain rule for time derivatives. Second, exact initial and boundary conditions (IC/BC) are imposed using approximate distance functions (ADFs) [Sukumar and Srivastava, CMAME, 2022]. However, ADFs introduce a trade-off as their second- or higher-order derivatives become unstable at the joining points of boundaries. To address this, we introduce an auxiliary network inspired by [Gladstone et al., NeurIPS ML4PS workshop, 2022]. Our method is validated through parameter extrapolation and interpolation, temporal extrapolation, and comparisons with analytical solutions.

math.NA↗

arXiv2024

SoMaSLAM: 2D Graph SLAM for Sparse Range Sensing with Soft Manhattan World Constraints

We propose a graph SLAM algorithm for sparse range sensing that incorporates a soft Manhattan world utilizing landmark-landmark constraints. Sparse range sensing is necessary for tiny robots that do not have the luxury of using heavy and expensive sensors. Existing SLAM methods dealing with sparse range sensing lack accuracy and accumulate drift error over time due to limited access to data points. Algorithms that cover this flaw using structural regularities, such as the Manhattan world (MW), have shortcomings when mapping real-world environments that do not coincide with the rules. We propose SoMaSLAM, a 2D graph SLAM designed for tiny robots with sparse range sensing. Our approach effectively maps sparse range data without enforcing strict structural regularities and maintains an adaptive graph. We implement the MW assumption as soft constraints, which we refer to as a soft Manhattan world. We propose novel soft landmark-landmark constraints to incorporate the soft MW into graph SLAM. Through extensive evaluation, we demonstrate that our proposed SoMaSLAM method improves localization accuracy on diverse datasets and is flexible enough to be used in the real world. We release our source code and sparse range datasets at https://SoMaSLAM.github.io/.

cs.RO↗

arXiv2024

Sampling low-fidelity outputs for estimation of high-fidelity density and its tails

In a multifidelity setting, data are available under the same conditions from two (or more) sources, e.g. computer codes, one being lower-fidelity but computationally cheaper, and the other higher-fidelity and more expensive. This work studies for which low-fidelity outputs, one should obtain high-fidelity outputs, if the goal is to estimate the probability density function of the latter, especially when it comes to the distribution tails and extremes. It is suggested to approach this problem from the perspective of the importance sampling of low-fidelity outputs according to some proposal distribution, combined with special considerations for the distribution tails based on extreme value theory. The notion of an optimal proposal distribution is introduced and investigated, in both theory and simulations. The approach is motivated and illustrated with an application to estimate the probability density function of record extremes of ship motions, obtained through two computer codes of different fidelities.

stat.ME↗

arXiv2024

Addressing Distribution Shift in RTB Markets via Exponential Tilting

In machine learning applications, distribution shifts between training and target environments can lead to significant drops in model performance. This study investigates the impact of such shifts on binary classification models within the Real-Time Bidding (RTB) market context, where selection bias contributes to these shifts. To address this challenge, we apply the Exponential Tilt Reweighting Alignment (ExTRA) algorithm, proposed by Maity et al. (2023). This algorithm estimates importance weights for the empirical risk by considering both covariate and label distributions, without requiring target label information, by assuming a specific weight structure. The goal of this study is to estimate weights that correct for the distribution shifts in RTB model and to evaluate the efficiency of the proposed model using simulated real-world data.

stat.ML↗

arXiv2024

Leveraging Temporal Contextualization for Video Action Recognition

We propose a novel framework for video understanding, called Temporally Contextualized CLIP (TC-CLIP), which leverages essential temporal information through global interactions in a spatio-temporal domain within a video. To be specific, we introduce Temporal Contextualization (TC), a layer-wise temporal information infusion mechanism for videos, which 1) extracts core information from each frame, 2) connects relevant information across frames for the summarization into context tokens, and 3) leverages the context tokens for feature encoding. Furthermore, the Video-conditional Prompting (VP) module processes context tokens to generate informative prompts in the text modality. Extensive experiments in zero-shot, few-shot, base-to-novel, and fully-supervised action recognition validate the effectiveness of our model. Ablation studies for TC and VP support our design choices. Our project page with the source code is available at https://github.com/naver-ai/tc-clip

cs.CV↗

arXiv2023

Interpretable Online Network Dictionary Learning for Inferring Long-Range Chromatin Interactions

Dictionary learning (DL) is commonly used in computational biology to tackle ubiquitous clustering problems due to its conceptual simplicity and relatively low computational complexity. However, DL algorithms produce results that lack interpretability and are not optimized for large-scale graph-structured data. We propose a novel DL algorithm called online convex network dictionary learning (online cvxNDL) that can handle extremely large datasets and enables the interpretation of dictionary elements, which serve as cluster representatives, through convex combinations of real measurements. Moreover, the algorithm can be applied to network-structured data via specialized subnetwork sampling techniques. To demonstrate the utility of our approach, we apply cvxNDL on 3D-genome RNAPII ChIA-Drop data to identify important long-range interaction patterns. ChIA-Drop probes higher-order interactions, and produces hypergraphs whose nodes represent genomic fragments. The hyperedges represent observed physical contacts. Our hypergraph model analysis creates an interpretable dictionary of long-range interaction patterns that accurately represent global chromatin physical contact maps. Using dictionary information, one can also associate the contact maps with RNA transcripts and infer cellular functions. Our results offer two key insights. First, we demonstrate that online cvxNDL retains the accuracy of classical DL methods while simultaneously ensuring unique interpretability and scalability. Second, we identify distinct collections of proximal and distal interaction patterns involving chromatin elements shared by related processes across different chromosomes, as well as patterns unique to specific chromosomes. To associate the dictionary elements with biological properties of the corresponding chromatin regions, we employ Gene Ontology enrichment analysis and perform RNA coexpression studies.

q-bio.GN↗

arXiv2023

Neural Collage Transfer: Artistic Reconstruction via Material Manipulation

Collage is a creative art form that uses diverse material scraps as a base unit to compose a single image. Although pixel-wise generation techniques can reproduce a target image in collage style, it is not a suitable method due to the solid stroke-by-stroke nature of the collage form. While some previous works for stroke-based rendering produced decent sketches and paintings, collages have received much less attention in research despite their popularity as a style. In this paper, we propose a method for learning to make collages via reinforcement learning without the need for demonstrations or collage artwork data. We design the collage Markov Decision Process (MDP), which allows the agent to handle various materials and propose a model-based soft actor-critic to mitigate the agent's training burden derived from the sophisticated dynamics of collage. Moreover, we devise additional techniques such as active material selection and complexity-based multi-scale collage to handle target images at any size and enhance the results' aesthetics by placing relatively more scraps in areas of high complexity. Experimental results show that the trained agent appropriately selected and pasted materials to regenerate the target image into a collage and obtained a higher evaluation score on content and style than pixel-wise generation methods. Code is available at https://github.com/northadventure/CollageRL.

cs.CV↗

arXiv2023

EXOT: Exit-aware Object Tracker for Safe Robotic Manipulation of Moving Object

Current robotic hand manipulation narrowly operates with objects in predictable positions in limited environments. Thus, when the location of the target object deviates severely from the expected location, a robot sometimes responds in an unexpected way, especially when it operates with a human. For safe robot operation, we propose the EXit-aware Object Tracker (EXOT) on a robot hand camera that recognizes an object's absence during manipulation. The robot decides whether to proceed by examining the tracker's bounding box output containing the target object. We adopt an out-of-distribution classifier for more accurate object recognition since trackers can mistrack a background as a target object. To the best of our knowledge, our method is the first approach of applying an out-of-distribution classification technique to a tracker output. We evaluate our method on the first-person video benchmark dataset, TREK-150, and on the custom dataset, RMOT-223, that we collect from the UR5e robot. Then we test our tracker on the UR5e robot in real-time with a conveyor-belt sushi task, to examine the tracker's ability to track target dishes and to determine the exit status. Our tracker shows 38% higher exit-aware performance than a baseline method. The dataset and the code will be released at https://github.com/hskAlena/EXOT.

cs.CV↗

arXiv2022

Towards Sequence-Level Training for Visual Tracking

Despite the extensive adoption of machine learning on the task of visual object tracking, recent learning-based approaches have largely overlooked the fact that visual tracking is a sequence-level task in its nature; they rely heavily on frame-level training, which inevitably induces inconsistency between training and testing in terms of both data distributions and task objectives. This work introduces a sequence-level training strategy for visual tracking based on reinforcement learning and discusses how a sequence-level design of data sampling, learning objectives, and data augmentation can improve the accuracy and robustness of tracking algorithms. Our experiments on standard benchmarks including LaSOT, TrackingNet, and GOT-10k demonstrate that four representative tracking models, SiamRPN++, SiamAttn, TransT, and TrDiMP, consistently improve by incorporating the proposed methods in training without modifying architectures.

cs.CV↗

arXiv2022

From Scratch to Sketch: Deep Decoupled Hierarchical Reinforcement Learning for Robotic Sketching Agent

We present an automated learning framework for a robotic sketching agent that is capable of learning stroke-based rendering and motor control simultaneously. We formulate the robotic sketching problem as a deep decoupled hierarchical reinforcement learning; two policies for stroke-based rendering and motor control are learned independently to achieve sub-tasks for drawing, and form a hierarchy when cooperating for real-world drawing. Without hand-crafted features, drawing sequences or trajectories, and inverse kinematics, the proposed method trains the robotic sketching agent from scratch. We performed experiments with a 6-DoF robot arm with 2F gripper to sketch doodles. Our experimental results show that the two policies successfully learned the sub-tasks and collaborated to sketch the target images. Also, the robustness and flexibility were examined by varying drawing tools and surfaces.

cs.RO↗

arXiv2022

Online Hybrid Lightweight Representations Learning: Its Application to Visual Tracking

This paper presents a novel hybrid representation learning framework for streaming data, where an image frame in a video is modeled by an ensemble of two distinct deep neural networks; one is a low-bit quantized network and the other is a lightweight full-precision network. The former learns coarse primary information with low cost while the latter conveys residual information for high fidelity to original representations. The proposed parallel architecture is effective to maintain complementary information since fixed-point arithmetic can be utilized in the quantized network and the lightweight model provides precise representations given by a compact channel-pruned network. We incorporate the hybrid representation technique into an online visual tracking task, where deep neural networks need to handle temporal variations of target appearances in real-time. Compared to the state-of-the-art real-time trackers based on conventional deep neural networks, our tracking algorithm demonstrates competitive accuracy on the standard benchmarks with a small fraction of computational cost and memory footprint.

cs.CV↗

arXiv2021

Rapid measurement of the local pressure amplitude in microchannel acoustophoresis using motile cells

Acoustic microfluidics (or acoustofluidics) provides a non-contact and label-free means to manipulate and interrogate bioparticles. Owing to their biocompatibility and precision, acoustofluidic approaches have enabled innovations in various areas of biomedical research. Future breakthroughs will rely on translation of these techniques from academic labs to clinical and industrial settings. Here, accurate characterization and standardization of device performance is crucial. Versatile, rapid, and widely accessible performance quantification is needed. We propose a field quantification method using motile Chlamydomonas reinhardtii algae cells. We previously reported qualitative mapping of acoustic fields using living microswimmers as active probes. In the present study, we extend our approach to achieve the challenging quantitative in situ measurement of the acoustic energy density. C. reinhardtii cells continuously swim in an imposed force field and dynamically redistribute as the field changes. This behavior allows accurate and complete, real-time performance monitoring, which can be easily applied and adopted within the acoustofluidics and broader microfluidics research communities. Additionally, the approach relies only on standard bright-field microscopy to assess the field under numerous conditions within minutes. We benchmark the method against conventional passive-particle tracking, achieving agreement within 1 % for field strengths from 0 to 100 J m-3 (0 to ~1 MPa).

physics.app-ph↗

arXiv2015

Three-photon-annihilation contributions to positronium energies at order $m α^7$

Positronium spectroscopy (n=1 hyperfine splitting, n=2 fine structure, and the 1S-2S interval) has reached a precision of order 1 MHz. Vigorous ongoing efforts to improve the experimental results motivate the calculation of the positronium energy levels at order $m α^7$. In this work we present the result for a complete class of such contributions--those involving virtual annihilation of positronium to three photons in an intermediate state. We find an energy shift of $2.6216(11) m α^7/(n π)^3$ for the spin-triplet S state with principal quantum number n.

hep-ph↗

arXiv2013

MCUIUC -- A New Framework for Metagenomic Read Compression

Metagenomics is an emerging field of molecular biology concerned with analyzing the genomes of environmental samples comprising many different diverse organisms. Given the nature of metagenomic data, one usually has to sequence the genomic material of all organisms in a batch, leading to a mix of reads coming from different DNA sequences. In deep high-throughput sequencing experiments, the volume of the raw reads is extremely high, frequently exceeding 600 Gb. With an ever increasing demand for storing such reads for future studies, the issue of efficient metagenomic compression becomes of paramount importance. We present the first known approach to metagenome read compression, termed MCUIUC (Metagenomic Compression at UIUC). The gist of the proposed algorithm is to perform classification of reads based on unique organism identifiers, followed by reference-based alignment of reads for individually identified organisms, and metagenomic assembly of unclassified reads. Once assembly and classification are completed, lossless reference based compression is performed via positional encoding. We evaluate the performance of the algorithm on moderate sized synthetic metagenomic samples involving 15 randomly selected organisms and describe future directions for improving the proposed compression method.

q-bio.GN↗

arXiv2013

MetaPar: Metagenomic Sequence Assembly via Iterative Reclassification

We introduce a parallel algorithmic architecture for metagenomic sequence assembly, termed MetaPar, which allows for significant reductions in assembly time and consequently enables the processing of large genomic datasets on computers with low memory usage. The gist of the approach is to iteratively perform read (re)classification based on phylogenetic marker genes and assembler outputs generated from random subsets of metagenomic reads. Once a sufficiently accurate classification within genera is performed, de novo metagenomic assemblers (such as Velvet or IDBA-UD) or reference based assemblers may be used for contig construction. We analyze the performance of MetaPar on synthetic data consisting of 15 randomly chosen species from the NCBI database through the effective gap and effective coverage metrics.

q-bio.QM↗