arXiv subjects

Hakho Lee

Publications and source records attributed to Hakho Lee.

7 recordsLinked to original sources

arXiv2023

Molecular fingerprinting of biological nanoparticles with a label-free optofluidic platform

Label-free detecting multiple analytes in a high-throughput fashion has been one of the long-sought goals in biosensing applications. Yet, for all-optical approaches, interfacing state-of-the-art label-free techniques with microfluidics tools that can process small volumes of sample with high throughput, and with surface chemistry that grants analyte specificity, poses a critical challenge to date. Here, we introduce an optofluidic platform that brings together state-of-the-art digital holography with PDMS microfluidics by using supported lipid bilayers as a surface chemistry building block to integrate both technologies. Specifically, this platform fingerprints heterogeneous biological nanoparticle populations via a multiplexed label-free immunoaffinity assay with single particle sensitivity. Herein, we first thoroughly characterise the robustness and performance of the platform, and then apply it to profile four distinct ovarian cell-derived extracellular vesicle populations over a panel of surface protein biomarkers, thus developing a unique biomarker fingerprint for each cell line. We foresee that our approach will find many applications where routine and multiplexed characterisation of biological nanoparticles is required.

q-bio.QM↗

arXiv2023

A Scalable Finite Difference Method for Deep Reinforcement Learning

Several low-bandwidth distributable black-box optimization algorithms in the family of finite differences such as Evolution Strategies have recently been shown to perform nearly as well as tailored Reinforcement Learning methods in some Reinforcement Learning domains. One shortcoming of these black-box methods is that they must collect information about the structure of the return function at every update, and can often employ only information drawn from a distribution centered around the current parameters. As a result, when these algorithms are distributed across many machines, a significant portion of total runtime may be spent with many machines idle, waiting for a final return and then for an update to be calculated. In this work we introduce a novel method to use older data in finite difference algorithms, which produces a scalable algorithm that avoids significant idle time or wasted computation.

cs.LG↗

arXiv2021

Agent Spaces

Exploration is one of the most important tasks in Reinforcement Learning, but it is not well-defined beyond finite problems in the Dynamic Programming paradigm (see Subsection 2.4). We provide a reinterpretation of exploration which can be applied to any online learning method. We come to this definition by approaching exploration from a new direction. After finding that concepts of exploration created to solve simple Markov decision processes with Dynamic Programming are no longer broadly applicable, we reexamine exploration. Instead of extending the ends of dynamic exploration procedures, we extend their means. That is, rather than repeatedly sampling every state-action pair possible in a process, we define the act of modifying an agent to itself be explorative. The resulting definition of exploration can be applied in infinite problems and non-dynamic learning methods, which the dynamic notion of exploration cannot tolerate. To understand the way that modifications of an agent affect learning, we describe a novel structure on the set of agents: a collection of distances (see footnote 7) $d_{a} \in A$, which represent the perspectives of each agent possible in the process. Using these distances, we define a topology and show that many important structures in Reinforcement Learning are well behaved under the topology induced by convergence in the agent space.

cs.AI↗

arXiv2020

Precise nanosizing with high dynamic range holography

Optical sensing is one of the key-enablers of modern diagnostics. Especially label-free imaging modalities hold great promise as they eliminate labeling procedures prior to analysis. However, scattering signals of nanometric particles scale with their volume-square. This unfavorable scaling makes it extremely difficult to quantitatively characterize intrinsically heterogeneous clinical samples, such as extracellular vesicles, as their signal variation easily exceeds the dynamic range of currently available cameras. Here, we introduce off-axis k-space holography that circumvents this limitation. By imaging the back-focal-plane of our microscope we project the scattering signal of all particles onto all camera pixels thus dramatically boosting the achievable dynamic range to up-to 110 dB. We validate our platform by detecting, and quantitatively sizing, metallic and dielectric particles over a 200x200 $μ$m field-of-view and demonstrate that independently performed signal calibrations allow correctly sizing particles made from different materials. Finally, we present quantitative size-distributions of extracellular vesicle samples.

physics.optics↗

arXiv2019

Evolution Strategies Converges to Finite Differences

Since the debut of Evolution Strategies (ES) as a tool for Reinforcement Learning by Salimans et al. 2017, there has been interest in determining the exact relationship between the Evolution Strategies gradient and the gradient of a similar class of algorithms, Finite Differences (FD).(Zhang et al. 2017, Lehman et al. 2018) Several investigations into the subject have been performed, investigating the formal motivational differences(Lehman et al. 2018) between ES and FD, as well as the differences in a standard benchmark problem in Machine Learning, the MNIST classification problem(Zhang et al. 2017). This paper proves that while the gradients are different, they converge as the dimension of the vector under optimization increases.

cs.NE↗

arXiv2018

High-throughput intensity diffraction tomography with a computational microscope

We demonstrate a motion-free intensity diffraction tomography technique that enables direct inversion of 3D phase and absorption from intensity-only measurements for weakly scattering samples. We derive a novel linear forward model, featuring slice-wise phase and absorption transfer functions using angled illumination. This new framework facilitates flexible and efficient data acquisition, enabling arbitrary sampling of the illumination angles. The reconstruction algorithm performs 3D synthetic aperture using a robust, computation and memory efficient slice-wise deconvolution to achieve resolution up to the incoherent limit. We demonstrate our technique with thick biological samples having both sparse 3D structures and dense cell clusters. We further investigate the limitation of our technique when imaging strongly scattering samples. Imaging performance and the influence of multiple scattering is evaluated using a 3D sample consisting of stacked phase and absorption resolution targets. This computational microscopy system is directly built on a standard commercial microscope with a simple LED array source add-on, and promises broad applications by leveraging the ubiquitous microscopy platforms with minimal hardware modifications.

eess.IV↗

arXiv2006

Integrated Cell Manipulation Systems

A new type of microfluidic system for biological cell manipulation, a CMOS/microfluidic hybrid, is demonstrated. The hybrid system starts with a custom-designed CMOS (complementary metal-oxide semiconductor) chip fabricated in a semiconductor foundry using standard integration circuit technology. A microfluidic channel is post-fabricated on top of the CMOS chip to provide biocompatible environment. The motion of individual biological cells that are tagged with magnetic beads is directly controlled by the CMOS chip that generates localized magnetic filed patterns using an on-chip array of micro-electromagnets. The speed and the programmability of the CMOS chip further allow for the dynamic reconfiguration of the magnetic fields, substantially increasing the manipulation capability of the hybrid system. The concept of a hybrid system is verified by simultaneously manipulating individual biological cells with microscopic resolution. A new operation protocol that exploits the fast speed of electronics to trap and move a large number of cells with less power consumption is also demonstrated. Combining the advantages of microelectronics, the CMOS/microfluidic hybrid approach presents a new model for a multifunctional lab-on-a chip for biological and medical applications.

q-bio.QM↗