Maxim S. Panov
YOU?
Author Swipe
View article: Noise Sources and Strategies for Signal Quality Improvement in Biological Imaging: A Review Focused on Calcium and Cell Membrane Voltage Imaging
Noise Sources and Strategies for Signal Quality Improvement in Biological Imaging: A Review Focused on Calcium and Cell Membrane Voltage Imaging Open
This review addresses the challenges of obtaining high-quality quantitative data in the optical imaging of membrane voltage and calcium dynamics. The paper provides a comprehensive overview and systematization of recent studies that analyz…
View article: Semirational Protein Engineering Yields Archaerhodopsin-3-Based Fluorescent Genetically Encoded Voltage Indicators with Enhanced Brightness and Red Shifted Absorption Bands
Semirational Protein Engineering Yields Archaerhodopsin-3-Based Fluorescent Genetically Encoded Voltage Indicators with Enhanced Brightness and Red Shifted Absorption Bands Open
A semirational engineering of novel bright variants of archaerhodopsin-based genetically encoded voltage indicators (GEVIs) is performed. The corresponding focused protein sequence library was generated using data derived from prior direct…
View article: Uncertainty-Aware Attention Heads: Efficient Unsupervised Uncertainty Quantification for LLMs
Uncertainty-Aware Attention Heads: Efficient Unsupervised Uncertainty Quantification for LLMs Open
Large language models (LLMs) exhibit impressive fluency, but often produce critical errors known as "hallucinations". Uncertainty quantification (UQ) methods are a promising tool for coping with this fundamental shortcoming. Yet, existing …
View article: Uncertainty-aware abstention in medical diagnosis based on medical texts
Uncertainty-aware abstention in medical diagnosis based on medical texts Open
This study addresses the critical issue of reliability for AI-assisted medical diagnosis. We focus on the selection prediction approach that allows the diagnosis system to abstain from providing the decision if it is not confident in the d…
View article: Rectifying Conformity Scores for Better Conditional Coverage
Rectifying Conformity Scores for Better Conditional Coverage Open
We present a new method for generating confidence sets within the split conformal prediction framework. Our method performs a trainable transformation of any given conformity score to improve conditional coverage while ensuring exact margi…
View article: Token-Level Density-Based Uncertainty Quantification Methods for Eliciting Truthfulness of Large Language Models
Token-Level Density-Based Uncertainty Quantification Methods for Eliciting Truthfulness of Large Language Models Open
Uncertainty quantification (UQ) is a prominent approach for eliciting truthful answers from large language models (LLMs). To date, information-based and consistency-based UQ have been the dominant UQ methods for text generation via LLMs. D…
View article: Efficient Diversity-based Experience Replay for Deep Reinforcement Learning
Efficient Diversity-based Experience Replay for Deep Reinforcement Learning Open
Experience replay is widely used to improve learning efficiency in reinforcement learning by leveraging past experiences. However, existing experience replay methods, whether based on uniform or prioritized sampling, often suffer from low …
View article: Probabilistic Conformal Prediction with Approximate Conditional Validity
Probabilistic Conformal Prediction with Approximate Conditional Validity Open
We develop a new method for generating prediction sets that combines the flexibility of conformal methods with an estimate of the conditional distribution $P_{Y \mid X}$. Existing methods, such as conformalized quantile regression and prob…
View article: Rational Design of Far-Red Archaerhodopsin-3-Based Fluorescent Genetically Encoded Voltage Indicators: from Elucidation of the Fluorescence Mechanism in Archers to Novel Red-Shifted Variants
Rational Design of Far-Red Archaerhodopsin-3-Based Fluorescent Genetically Encoded Voltage Indicators: from Elucidation of the Fluorescence Mechanism in Archers to Novel Red-Shifted Variants Open
Genetically encoded voltage indicators (GEVIs) have found wide applications as molecular tools for visualization of changes in cell membrane potential. Among others, several classes of archaerhodopsin-3-based GEVIs have been developed and …
View article: Generalization error of spectral algorithms
Generalization error of spectral algorithms Open
The asymptotically precise estimation of the generalization of kernel methods has recently received attention due to the parallels between neural networks and their associated kernels. However, prior works derive such estimates for trainin…
View article: Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantification
Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantification Open
Large language models (LLMs) are notorious for hallucinating, i.e., producing erroneous claims in their output. Such hallucinations can be dangerous, as occasional factual inaccuracies in the generated text might be obscured by the rest of…
View article: From Risk to Uncertainty: Generating Predictive Uncertainty Measures via Bayesian Estimation
From Risk to Uncertainty: Generating Predictive Uncertainty Measures via Bayesian Estimation Open
There are various measures of predictive uncertainty in the literature, but their relationships to each other remain unclear. This paper uses a decomposition of statistical pointwise risk into components, associated with different sources …
View article: Efficient Conformal Prediction under Data Heterogeneity
Efficient Conformal Prediction under Data Heterogeneity Open
Conformal Prediction (CP) stands out as a robust framework for uncertainty quantification, which is crucial for ensuring the reliability of predictions. However, common CP methods heavily rely on data exchangeability, a condition often vio…
View article: Dirichlet-based Uncertainty Quantification for Personalized Federated Learning with Improved Posterior Networks
Dirichlet-based Uncertainty Quantification for Personalized Federated Learning with Improved Posterior Networks Open
In modern federated learning, one of the main challenges is to account for inherent heterogeneity and the diverse nature of data distributions for different clients. This problem is often addressed by introducing personalization of the mod…
View article: Fluorescence of the Retinal Chromophore in Microbial and Animal Rhodopsins
Fluorescence of the Retinal Chromophore in Microbial and Animal Rhodopsins Open
Fluorescence of the vast majority of natural opsin-based photoactive proteins is extremely low, in accordance with their functions that depend on efficient transduction of absorbed light energy. However, several recently proposed classes o…
View article: Simultaneous Catechol and Hydroquinone Detection with Laser Fabricated MOF-Derived Cu-CuO@C Composite Electrochemical Sensor
Simultaneous Catechol and Hydroquinone Detection with Laser Fabricated MOF-Derived Cu-CuO@C Composite Electrochemical Sensor Open
The conversion of metal-organic frameworks (MOFs) into advanced functional materials offers a promising route for producing unique nanomaterials. MOF-derived systems have the potential to overcome the drawbacks of MOFs, such as low electri…
View article: Ruthenium-Based Electrode Modified by Gold Particles as Voltammetric Sensor for Non-Enzymatic Epinephrine Detection
Ruthenium-Based Electrode Modified by Gold Particles as Voltammetric Sensor for Non-Enzymatic Epinephrine Detection Open
A simple and efficient technique for voltammetric enzymeless detection of epinephrine (EP) is proposed. The technique applies hierarchical Ru-based electrodes modified by Au nano- to micro-sized particles, which were produced with laser-as…
View article: Fluorescence of the retinal chromophore in microbial and animal rhodopsins
Fluorescence of the retinal chromophore in microbial and animal rhodopsins Open
Fluorescence of the vast majority of natural opsin-based photoactive proteins is extremely low in accordance with their functions that depend on efficient transduction of absorbed light energy. However, recently proposed several classes of…
View article: LM-Polygraph: Uncertainty Estimation for Language Models
LM-Polygraph: Uncertainty Estimation for Language Models Open
Recent advancements in the capabilities of large language models (LLMs) have paved the way for a myriad of groundbreaking applications in various fields. However, a significant challenge arises as these models often "hallucinate", i.e., fa…
View article: Conformal Prediction for Federated Uncertainty Quantification Under Label Shift
Conformal Prediction for Federated Uncertainty Quantification Under Label Shift Open
Federated Learning (FL) is a machine learning framework where many clients collaboratively train models while keeping the training data decentralized. Despite recent advances in FL, the uncertainty quantification topic (UQ) remains partial…
View article: Rapid and effective method of laser metallization of dielectric materials using deep eutectic solvents with copper acetate
Rapid and effective method of laser metallization of dielectric materials using deep eutectic solvents with copper acetate Open
In this paper, we proposed rapid laser-induced synthesis of copper micropatterns on the surface of oxide glass from deep eutectic solvents (DESs) consisting of choline chloride, citric or tartaric acid and copper acetate as copper plating …
View article: Effect of Gd3+, La3+, Lu3+ Co-Doping on the Morphology and Luminescent Properties of NaYF4:Sm3+ Phosphors
Effect of Gd3+, La3+, Lu3+ Co-Doping on the Morphology and Luminescent Properties of NaYF4:Sm3+ Phosphors Open
The series of luminescent NaYF4:Sm3+ nano- and microcrystalline materials co-doped by La3+, Gd3+, and Lu3+ ions were synthesized by hydrothermal method using rare earth chlorides as the precursors and citric acid as a stabilizing agent. Th…
View article: Fluorescence Imaging of Cell Membrane Potential: From Relative Changes to Absolute Values
Fluorescence Imaging of Cell Membrane Potential: From Relative Changes to Absolute Values Open
Membrane potential is a fundamental property of biological cells. Changes in membrane potential characterize a vast number of vital biological processes, such as the activity of neurons and cardiomyocytes, tumorogenesis, cell-cycle progres…
View article: Lipid Alteration Signature in the Blood Plasma of Individuals With Schizophrenia, Depression, and Bipolar Disorder
Lipid Alteration Signature in the Blood Plasma of Individuals With Schizophrenia, Depression, and Bipolar Disorder Open
Importance No clinically applicable diagnostic test exists for severe mental disorders. Lipids harbor potential as disease markers. Objective To define a reproducible profile of lipid alterations in the blood plasma of patients with schizo…
View article: Scalable Batch Acquisition for Deep Bayesian Active Learning
Scalable Batch Acquisition for Deep Bayesian Active Learning Open
In deep active learning, it is especially important to choose multiple examples to markup at each step to work efficiently, especially on large datasets. At the same time, existing solutions to this problem in the Bayesian setup, such as B…
View article: Active Learning for Abstractive Text Summarization
Active Learning for Abstractive Text Summarization Open
Construction of human-curated annotated datasets for abstractive text summarization (ATS) is very time-consuming and expensive because creating each instance requires a human annotator to read a long document and compose a shorter summary …