arXiv subjects

Sujin Lee

Publications and source records attributed to Sujin Lee.

16 recordsLinked to original sources

arXiv2026

EvoStruct: Bridging Evolutionary and Structural Priors for Antibody CDR Design via Protein Language Model Adaptation

Equivariant graph neural network (GNN) methods for antibody complementarity-determining region (CDR) design achieve the highest sequence recovery but suffer from severe vocabulary collapse. The current best GNN methods over-predict very few amino acids, such as tyrosine and glycine, while ignoring functionally important residues. We trace this failure to GNN encoders learning amino acid distributions de novo from limited structural data, discarding substitution patterns encoded in evolutionary databases. To resolve this, we propose EvoStruct, which bridges a frozen protein language model (PLM) with 3D structural context from an E(3)-equivariant GNN via a cross-attention adapter. Unlike prior PLM-structure adapters for general protein design, EvoStruct targets the vocabulary collapse problem specific to CDR design through progressive PLM unfreezing and R-Drop consistency regularization. On the CHIMERA-Bench dataset, EvoStruct achieves the highest amino acid recovery and lowest perplexity among several antibody design methods, improving sequence recovery by 16% and reducing perplexity by 43% relative to the best GNN baselines, while recovering 2.3x greater amino acid diversity and the highest binding-pair correlation with ground truth.

cs.LG↗

arXiv2026

ConTact: Contact-First Antibody CDR Design via Explicit Interface Reasoning

Computational antibody CDR design methods condition on antigen structure to generate binding loops. Yet, the existing architectures conflate two fundamentally distinct sub-problems: identifying which CDR positions will contact the antigen, and selecting amino acids at those positions. This forces models to learn contact reasoning implicitly through uniform message passing, diluting antigen signal across all positions equally. We introduce ConTact, a contact-then-act architecture that explicitly decomposes CDR design into three cascaded stages: learning surface complementarity fingerprints, predicting CDR-antigen contacts, and injecting contact-gated antigen features into the prediction head. A distance-biased cross-attention module encodes geometric priors favoring spatial neighbors, a contact-weighted cross-entropy loss concentrates gradient signal on binding-critical positions, and a standoff-aware contact-recovery loss places the designed loop against its native epitope. On the ChiMERa-Bench dataset, ConTact surpasses the baselines on the binding interface and structural metrics (backbone RMSD, fraction of native contacts, interface RMSD, DockQ, and epitope F1) on all three splits, improving native-contact recovery by 11 to 16%, while matching the best sequence recovery. The source code is available at: https://github.com/mansoor181/ConTact.git

cs.LG↗

arXiv2026

Cooperative Informative Sensing for Monitoring Dynamic Indoor Environments via Multi-Agent Reinforcement Learning

Monitoring human activity in indoor environments is important for applications such as facility management, safety assessment, and space utilization analysis. While mobile robot teams offer the potential to actively improve observation quality, existing multi-robot monitoring and active perception approaches typically rely on coverage or visitation based objectives that are weakly aligned with the accuracy requirements of human-centric monitoring tasks. In this work, we formulate cooperative active observation as a decentralized control problem in which multiple robots adjust their motion to directly optimize monitoring accuracy under partial observability. We propose a learning-based framework for cooperative policies from decentralized observations using multi-agent reinforcement learning (MARL), supported by an architecture that handles variable numbers of humans and temporal dependencies. Simulation results across diverse indoor environments and monitoring tasks show that the proposed approach consistently outperforms classical coverage, persistent monitoring, and learning-free multi-robot baselines, while remaining robust to changes in the number of observed humans.

cs.RO↗

arXiv2024

Exciton Fission Enhanced Silicon Solar Cell

While silicon solar cells dominate global photovoltaic energy production, their continued improvement is hindered by the single junction limit. One potential solution is to use molecular singlet exciton fission to generate two electrons from each absorbed high-energy photon. We demonstrate that the long-standing challenge of coupling molecular excited states to silicon solar cells can be overcome using sequential charge transfer. Combining zinc phthalocyanine, aluminum oxide, and a shallow junction crystalline silicon microwire solar cell, the peak charge generation efficiency per photon absorbed in tetracene is (138 +- 6)%, comfortably surpassing the quantum efficiency limit for conventional silicon solar cells and establishing a new, scalable approach to low cost, high efficiency photovoltaics.

cond-mat.mtrl-sci↗

arXiv2024

An Application of Large Language Models to Coding Negotiation Transcripts

In recent years, Large Language Models (LLM) have demonstrated impressive capabilities in the field of natural language processing (NLP). This paper explores the application of LLMs in negotiation transcript analysis by the Vanderbilt AI Negotiation Lab. Starting in September 2022, we applied multiple strategies using LLMs from zero shot learning to fine tuning models to in-context learning). The final strategy we developed is explained, along with how to access and use the model. This study provides a sense of both the opportunities and roadblocks for the implementation of LLMs in real life applications and offers a model for how LLMs can be applied to coding in other fields.

cs.CL↗

arXiv2024

Complex zeros of Bessel function derivatives and associated orthogonal polynomials

We introduce a sequence of orthogonal polynomials whose associated moments are the Rayleigh-type sums, involving the zeros of the Bessel derivative $J_\nu'$ of order $\nu$. We also discuss the fundamental properties of those polynomials such as recurrence, orthogonality, etc. Consequently, we obtain a formula for the Hankel determinant, elements of which are chosen as the aforementioned Rayleigh-type sums. As an application, we complete the Hurwitz-type theorem for $J_\nu'$, which deals with the number of complex zeros of $J_\nu'$ depending on the range of $\nu$.

math.CA↗

arXiv2023

On the generalized interlacing property for the zeros of Bessel functions

This paper investigates a generalized interlacing property between Bessel functions, particularly $J_\nu$ and $J_\mu$, where the difference $|\nu-\mu|$ exceeds $2$. This interlacing phenomenon is marked by a compensatory interaction with the zeros of Lommel polynomials, extending our understanding beyond the traditional $|\nu-\mu| \le 2$ regime. The paper extends the generalized interlacing property to cylinder functions and derivative of Bessel functions, as an application. It is also discussed that Siegel's extension of Bourget hypothesis to rational numbers of $\nu$ cannot be further improved to arbitrary real numbers.

math.CA↗

arXiv2023

Atom-level design strategy for hydrogen evolution reaction of transition metal dichalcogenides catalysts

Two-dimensional transition metal dichalcogenides are among the most promising materials for water-splitting catalysts. While a variety of methods have been applied to promote the hydrogen evolution reaction on the transition metal dichalcogenides, doping of transition metal heteroatoms have attracted much attention since it provides effective ways to optimize the hydrogen adsorption and H2 generation reactions. Herein, we provide in-depth and systematic analyses on the trends of the free energy of hydrogen adsorption (ΔGH*), the most well-known descriptor for evaluating hydrogen evolution reaction performance, in the doped transition metal dichalcogenides. Using the total 150 doped transition metal dichalcogenides, we carried out the atom-level analysis on the origin of ΔGH* changes upon the transition metal heteroatom doping, and suggest two key factors that govern the hydrogen adsorption process on the doped transition metal dichalcogenides: 1) the changes in the charge of chalcogen atoms where hydrogen atoms adsorbed for the early transition metal doped structures, and 2) the structural deformation energies accompanying in introduced dopants for the late transition metal doped structures. Based on our findings, we interpret from a new perspective how vacancies in the TM-doped TMDs can provide optimal ΔGH* in HER. We suggest electrostatic control for early TM doped systems and structural control for late TM doped systems as the effective strategies for the thermoneutral ΔGH* in TMD.

cond-mat.mtrl-sci↗

arXiv2022

Bulk Assembly of Organic Metal Halide Nanoribbons

Organic metal halide hybrids with low-dimensional structures at the molecular level have received great attention recently for their exceptional structural tunability and unique photophysical properties. Here we report for the first time the synthesis and characterization of a one-dimensional (1D) organic metal halide hybrid material, which contains metal halide nanoribbons with a width of three octahedral units. It is found that this material with a chemical formula C$_8$H$_{28}$N$_5$Pb$_3$Cl$_{11}$ shows a dual emission with a photoluminescence quantum efficiency (PLQE) of around 25% under ultraviolet (UV) light irradiation. Photophysical studies and density functional theory (DFT) calculations suggest the coexisting of delocalized free excitons and localized self-trapped excitons in metal halide nanoribbons leading to the dual emission. This work shows once again the exceptional tunability of organic metal halide hybrids that bridge between molecular systems with localized states and crystalline ones with electronic bands.

cond-mat.mtrl-sci↗

arXiv2021

Configurations of rectangles in a set in $\mathbb{F}_q^2$

Let $\mathbb{F}_q$ be a finite field of order $q$. In this paper, we study the distribution of rectangles in a given set in $\mathbb{F}_q^2$. More precisely, for any $0<δ\le 1$, we prove that there exists an integer $q_0=q_0(δ)$ with the following property: if $q\ge q_0$ and $A$ is a multiplicative subgroup of $\mathbb{F}^*_q$ with $|A|\ge q^{2/3}$, then any set $S\subset \mathbb{F}_q^2$ with $|S|\ge δq^2$ contains at least $\gg \frac{|S|^4|A|^2}{q^5}$ rectangles with side-lengths in $A$. We also consider the case of rectangles with one fixed side-length and the other in a multiplicative subgroup $A$.

math.CO↗

arXiv2021

Image Compression with Recurrent Neural Network and Generalized Divisive Normalization

Image compression is a method to remove spatial redundancy between adjacent pixels and reconstruct a high-quality image. In the past few years, deep learning has gained huge attention from the research community and produced promising image reconstruction results. Therefore, recent methods focused on developing deeper and more complex networks, which significantly increased network complexity. In this paper, two effective novel blocks are developed: analysis and synthesis block that employs the convolution layer and Generalized Divisive Normalization (GDN) in the variable-rate encoder and decoder side. Our network utilizes a pixel RNN approach for quantization. Furthermore, to improve the whole network, we encode a residual image using LSTM cells to reduce unnecessary information. Experimental results demonstrated that the proposed variable-rate framework with novel blocks outperforms existing methods and standard image codecs, such as George's ~\cite{002} and JPEG in terms of image similarity. The project page along with code and models are available at https://khawar512.github.io/cvpr/

eess.IV↗

arXiv2021

On the finite field cone restriction conjecture in four dimensions and applications in incidence geometry

The first purpose of this paper is to solve completely the finite field cone restriction conjecture in four dimensions with $-1$ non-square. The second is to introduce a new approach to study incidence problems via restriction theory. More precisely, using the cone restriction estimates, we will prove sharp point-sphere incidence bounds associated with complex-valued functions for sphere sets of small size. Our incidence bounds with a specific function improve significantly a result given by Cilleruelo, Iosevich, Lund, Roche-Newton, and Rudnev.

math.CA↗

arXiv2021

Averages and maximal averages over Product j-varieties in finite fields

We study both averaging and maximal averaging problems for Product $j$-varieties defined by $Π_j=\{x\in \mathbb F_q^d: \prod_{k=1}^d x_k=j\}$ for $j\in \mathbb F_q^*,$ where $\mathbb F_q^d$ denotes a $d$-dimensional vector space over the finite field $\mathbb F_q$ with $q$ elements. We prove the sharp $L^p\to L^r$ boundedness of averaging operators associated to Product $j$-varieties. We also obtain the optimal $L^p$ estimate for a maximal averaging operator related to a family of Product $j$-varieties $\{Π_j\}_{j\in \mathbb F_q^*}.$

math.CA↗

arXiv2020

On restriction estimates for the zero radius sphere over finite fields

In this paper, we solve completely the $L^2\to L^r$ extension conjecture for the zero radius sphere over finite fields. We also obtain the sharp $L^p\to L^4$ extension estimate for non-zero radii spheres over finite fields, which improves significantly the previous result of the first and second listed authors.

math.CA↗

arXiv2018

Sum-product estimates over arbitrary finite fields

In this paper we prove some results on sum-product estimates over arbitrary finite fields. More precisely, we show that for sufficiently small sets $A\subset \mathbb{F}_q$ we have \[|(A-A)^2+(A-A)^2|\gg |A|^{1+\frac{1}{21}}.\] This can be viewed as the Erdős distinct distances problem for Cartesian product sets over arbitrary finite fields. We also prove that \[\max\{|A+A|, |A^2+A^2|\}\gg |A|^{1+\frac{1}{42}}, ~|A+A^2|\gg |A|^{1+\frac{1}{84}}.\]

math.NT↗

arXiv2016

Informative Path Planning and Mapping with Multiple UAVs in Wind Fields

Informative path planning (IPP) is used to design paths for robotic sensor platforms to extract the best/maximum possible information about a quantity of interest while operating under a set of constraints, such as the dynamic feasibility of vehicles. The key challenges of IPP are the strong coupling in multiple layers of decisions: the selection of locations to visit, the allocation of sensor platforms to those locations; and the processing of the gathered information along the paths. This paper presents a systematic procedure for IPP and environmental mapping using multiple UAV sensor platforms. It (a) selects the best locations to observe, (b) calculates the cost and finds the best paths for each UAV, and (c) estimates the measurement value within a given region using the Gaussian process (GP) regression framework. An illustrative example of RF intensity field mapping is presented to demonstrate the validity and applicability of the proposed approach.

cs.RO↗