arXiv subjects

Sung Kim

Publications and source records attributed to Sung Kim.

11 recordsLinked to original sources

arXiv2025

Full System Architecture Modeling for Wearable Egocentric Contextual AI

The next generation of human-oriented computing will require always-on, spatially-aware wearable devices to capture egocentric vision and functional primitives (e.g., Where am I? What am I looking at?, etc.). These devices will sense an egocentric view of the world around us to observe all human-relevant signals across space and time to construct and maintain a user's personal context. This personal context, combined with advanced generative AI, will unlock a powerful new generation of contextual AI personal assistants and applications. However, designing a wearable system to support contextual AI is a daunting task because of the system's complexity and stringent power constraints due to weight and battery restrictions. To understand how to guide design for such systems, this work provides the first complete system architecture view of one such wearable contextual AI system (Aria2), along with the lessons we have learned through the system modeling and design space exploration process. We show that an end-to-end full system model view of such systems is vitally important, as no single component or category overwhelmingly dominates system power. This means long-range design decisions and power optimizations need to be made in the full system context to avoid running into limits caused by other system bottlenecks (i.e., Amdahl's law as applied to power) or as bottlenecks change. Finally, we reflect on lessons and insights for the road ahead, which will be important toward eventually enabling all-day, wearable, contextual AI systems.

cs.AR↗

arXiv2025

Fusion Trees and Homological Representations

We establish an identification between the spaces of $\alpha$-fusion trees in non-semisimple topological quantum computation (NSS TQC) and a family of homological representations of the braid group known as the Lawrence representations specialized at roots of unity. Leveraging this connection, we provide a new proof of Ito's colored Alexander invariant formula using graphical calculus. Inspired by Anghel's topological model, we derive a formula involving the Hermitian pairing of fusion trees. This formula verifies that non-semisimple quantum knot invariants can be explicitly encoded via the language of fusion trees in the NSS TQC mathematical architecture.

math.GT↗

arXiv2025

Robust Universal Braiding with Non-semisimple Ising Anyons

Non-semisimple extensions of the Ising anyon model developed in our previous work enable universal topological quantum computation via braiding alone, overcoming the Clifford-only limitation of semisimple theories. The non-semisimple theory provides new anyon types indexed by a real parameter $\alpha$, the neglecton. Braiding acts unitarily with respect to an indefinite Hermitian form, while the computational subspace sits in a positive-definite sector. We demonstrate that this universality is robust, persisting over an open interval of the neglecton parameter $\alpha$ where the computational subspace remains positive-definite. We identify special values of $\alpha$ where the physical subspace decouples exactly from negative-norm components, ensuring fully unitary evolution and suppressed leakage. We further present an alternative encoding supporting exact single-qubit Clifford gates alongside a non-Clifford phase gate. We show that high-precision tuning of $\alpha$ is not required for efficient gate compilation, significantly enhancing the physical plausibility of non-semisimple anyonic architectures.

quant-ph↗

arXiv2024

Universal quantum computation using Ising anyons from a non-semisimple Topological Quantum Field Theory

We propose a framework for topological quantum computation using newly discovered non-semisimple analogs of topological quantum field theories in 2+1 dimensions. These enhanced theories offer more powerful models for quantum computation. The conventional theory of Ising anyons, which is believed to describe excitations in the $\nu = 5/2$ fractional quantum Hall state, is not universal for quantum computation via braiding of quasiparticles. However, we show that the non-semisimple theory introduces new anyon types that extend the Ising framework. By adding just one new anyon type, universal quantum computation can be achieved through braiding alone. This result opens new avenues for realizing fault-tolerant quantum computing in topologically ordered systems.

quant-ph↗

arXiv2023

A Multimodal Approach for Advanced Pest Detection and Classification

This paper presents a novel multi modal deep learning framework for enhanced agricultural pest detection, combining tiny-BERT's natural language processing with R-CNN and ResNet-18's image processing. Addressing limitations of traditional CNN-based visual methods, this approach integrates textual context for more accurate pest identification. The R-CNN and ResNet-18 integration tackles deep CNN issues like vanishing gradients, while tiny-BERT ensures computational efficiency. Employing ensemble learning with linear regression and random forest models, the framework demonstrates superior discriminate ability, as shown in ROC and AUC analyses. This multi modal approach, blending text and image data, significantly boosts pest detection in agriculture. The study highlights the potential of multi modal deep learning in complex real-world scenarios, suggesting future expansions in diversity of datasets, advanced data augmentation, and cross-modal attention mechanisms to enhance model performance.

cs.CV↗

arXiv2021

Versa: A Dataflow-Centric Multiprocessor with 36 Systolic ARM Cortex-M4F Cores and a Reconfigurable Crossbar-Memory Hierarchy in 28nm

We present Versa, an energy-efficient processor with 36 systolic ARM Cortex-M4F cores and a runtime-reconfigurable memory hierarchy. Versa exploits algorithm-specific characteristics in order to optimize bandwidth, access latency, and data reuse. Measured on a set of kernels with diverse data access, control, and synchronization characteristics, reconfiguration between different Versa modes yields median energy-efficiency improvements of 11.6x and 37.2x over mobile CPU and GPU baselines, respectively.

cs.AR↗

arXiv2021

Zesting produces modular isotopes and explains their topological invariants

We show that the ribbon zesting construction can produce modular isotopes -- different modular fusion categories with the same modular data. The result relies on the observation that the Reshetikhin-Turaev invariants of framed links associated to a ribbon fusion category satisfy a factorization property under zesting. This gives a new perspective on using topological invariants to classify topological order in light of modular data not being a complete invariant.

math.QA↗

arXiv2019

Bandwidth Extension on Raw Audio via Generative Adversarial Networks

Neural network-based methods have recently demonstrated state-of-the-art results on image synthesis and super-resolution tasks, in particular by using variants of generative adversarial networks (GANs) with supervised feature losses. Nevertheless, previous feature loss formulations rely on the availability of large auxiliary classifier networks, and labeled datasets that enable such classifiers to be trained. Furthermore, there has been comparatively little work to explore the applicability of GAN-based methods to domains other than images and video. In this work we explore a GAN-based method for audio processing, and develop a convolutional neural network architecture to perform audio super-resolution. In addition to several new architectural building blocks for audio processing, a key component of our approach is the use of an autoencoder-based loss that enables training in the GAN framework, with feature losses derived from unlabeled data. We explore the impact of our architectural choices, and demonstrate significant improvements over previous works in terms of both objective and perceptual quality.

cs.SD↗

arXiv2018

MATIC: Learning Around Errors for Efficient Low-Voltage Neural Network Accelerators

As a result of the increasing demand for deep neural network (DNN)-based services, efforts to develop dedicated hardware accelerators for DNNs are growing rapidly. However,while accelerators with high performance and efficiency on convolutional deep neural networks (Conv-DNNs) have been developed, less progress has been made with regards to fully-connected DNNs (FC-DNNs). In this paper, we propose MATIC (Memory Adaptive Training with In-situ Canaries), a methodology that enables aggressive voltage scaling of accelerator weight memories to improve the energy-efficiency of DNN accelerators. To enable accurate operation with voltage overscaling, MATIC combines the characteristics of destructive SRAM reads with the error resilience of neural networks in a memory-adaptive training process. Furthermore, PVT-related voltage margins are eliminated using bit-cells from synaptic weights as in-situ canaries to track runtime environmental variation. Demonstrated on a low-power DNN accelerator that we fabricate in 65 nm CMOS, MATIC enables up to 60-80 mV of voltage overscaling (3.3x total energy reduction versus the nominal voltage), or 18.6x application error reduction.

cs.NE↗

arXiv2017

Exploring Computation-Communication Tradeoffs in Camera Systems

Cameras are the defacto sensor. The growing demand for real-time and low-power computer vision, coupled with trends towards high-efficiency heterogeneous systems, has given rise to a wide range of image processing acceleration techniques at the camera node and in the cloud. In this paper, we characterize two novel camera systems that use acceleration techniques to push the extremes of energy and performance scaling, and explore the computation-communication tradeoffs in their design. The first case study targets a camera system designed to detect and authenticate individual faces, running solely on energy harvested from RFID readers. We design a multi-accelerator SoC design operating in the sub-mW range, and evaluate it with real-world workloads to show performance and energy efficiency improvements over a general purpose microprocessor. The second camera system supports a 16-camera rig processing over 32 Gb/s of data to produce real-time 3D-360 degree virtual reality video. We design a multi-FPGA processing pipeline that outperforms CPU and GPU configurations by up to 10x in computation time, producing panoramic stereo video directly from the camera rig at 30 frames per second. We find that an early data reduction step, either before complex processing or offloading, is the most critical optimization for in-camera systems.

cs.AR↗

arXiv2010

Boundary Effects on the Determination of Metamaterial Parameters from Normal Incidence Reflection and Transmission Measurements

A method is described for the determination of the effective electromagnetic parameters of a metamaterial based only on external measurements or simulations, taking boundary effects at the interfaces between a conventional material and metamaterial into account. Plane-wave reflection and transmission coefficients at the interfaces are regarded as additional unknowns to be determined, rather than explicitly dependent on the material parameters. Our technique is thus analogous to the line-reflect-line (LRL) calibration method in microwave measurements. The refractive index can be determined from S-parameters for two samples of different thickness. The effective wave impedance requires the additional assumption that generalized sheet transition conditions (GSTCs) account for the boundary effects. Expressions for the bulk permittivity and permeability then follow easily. Our method is validated by comparison with the results using the Nicolson-Ross-Weir (NRW) for determining properties of an ordinary material measured in a coaxial line. Utilizing S-parameters obtained from 3-D full wave simulations, we test the method on magnetodielectric metamaterials. We compare the results from our method and the conventional one that does not consider boundary effects. Moreover, it is shown that results from our method are consistent under changes in reference plane location, whereas the results from other methods are not.

physics.class-ph↗