Showing 1–50 of 100 results for author: Kuznetsov, I

Systematic Task Exploration with LLMs: A Study in Citation Text Generation

Authors: Furkan Şahinuç, Ilia Kuznetsov, Yufang Hou, Iryna Gurevych

Abstract: Large language models (LLMs) bring unprecedented flexibility in defining and executing complex, creative natural language generation (NLG) tasks. Yet, this flexibility brings new challenges, as it introduces new degrees of freedom in formulating the task inputs and instructions and in evaluating model performance. To facilitate the exploration of creative NLG tasks, we propose a three-component re… ▽ More Large language models (LLMs) bring unprecedented flexibility in defining and executing complex, creative natural language generation (NLG) tasks. Yet, this flexibility brings new challenges, as it introduces new degrees of freedom in formulating the task inputs and instructions and in evaluating model performance. To facilitate the exploration of creative NLG tasks, we propose a three-component research framework that consists of systematic input manipulation, reference data, and output measurement. We use this framework to explore citation text generation -- a popular scholarly NLP task that lacks consensus on the task definition and evaluation metric and has not yet been tackled within the LLM paradigm. Our results highlight the importance of systematically investigating both task instruction and input configuration when prompting LLMs, and reveal non-trivial relationships between different evaluation metrics used for citation text generation. Additional human generation and human evaluation experiments provide new qualitative insights into the task to guide future research in citation text generation. We make our code and data publicly available. △ Less

Submitted 4 July, 2024; originally announced July 2024.

Comments: Accepted to ACL 2024 (Main)

M2QA: Multi-domain Multilingual Question Answering

Authors: Leon Engländer, Hannah Sterz, Clifton Poth, Jonas Pfeiffer, Ilia Kuznetsov, Iryna Gurevych

Abstract: Generalization and robustness to input variation are core desiderata of machine learning research. Language varies along several axes, most importantly, language instance (e.g. French) and domain (e.g. news). While adapting NLP models to new languages within a single domain, or to new domains within a single language, is widely studied, research in joint adaptation is hampered by the lack of evalu… ▽ More Generalization and robustness to input variation are core desiderata of machine learning research. Language varies along several axes, most importantly, language instance (e.g. French) and domain (e.g. news). While adapting NLP models to new languages within a single domain, or to new domains within a single language, is widely studied, research in joint adaptation is hampered by the lack of evaluation datasets. This prevents the transfer of NLP systems from well-resourced languages and domains to non-dominant language-domain combinations. To address this gap, we introduce M2QA, a multi-domain multilingual question answering benchmark. M2QA includes 13,500 SQuAD 2.0-style question-answer instances in German, Turkish, and Chinese for the domains of product reviews, news, and creative writing. We use M2QA to explore cross-lingual cross-domain performance of fine-tuned models and state-of-the-art LLMs and investigate modular approaches to domain and language adaptation. We witness 1) considerable performance variations across domain-language combinations within model classes and 2) considerable performance drops between source and target language-domain combinations across all model sizes. We demonstrate that M2QA is far from solved, and new methods to effectively transfer both linguistic and domain-specific information are necessary. We make M2QA publicly available at https://github.com/UKPLab/m2qa. △ Less

Submitted 1 July, 2024; originally announced July 2024.

Re3: A Holistic Framework and Dataset for Modeling Collaborative Document Revision

Authors: Qian Ruan, Ilia Kuznetsov, Iryna Gurevych

Abstract: Collaborative review and revision of textual documents is the core of knowledge work and a promising target for empirical analysis and NLP assistance. Yet, a holistic framework that would allow modeling complex relationships between document revisions, reviews and author responses is lacking. To address this gap, we introduce Re3, a framework for joint analysis of collaborative document revision.… ▽ More Collaborative review and revision of textual documents is the core of knowledge work and a promising target for empirical analysis and NLP assistance. Yet, a holistic framework that would allow modeling complex relationships between document revisions, reviews and author responses is lacking. To address this gap, we introduce Re3, a framework for joint analysis of collaborative document revision. We instantiate this framework in the scholarly domain, and present Re3-Sci, a large corpus of aligned scientific paper revisions manually labeled according to their action and intent, and supplemented with the respective peer reviews and human-written edit summaries. We use the new data to provide first empirical insights into collaborative document revision in the academic domain, and to assess the capabilities of state-of-the-art LLMs at automating edit analysis and facilitating text-based collaboration. We make our annotation environment and protocols, the resulting data and experimental code publicly available. △ Less

Submitted 31 May, 2024; originally announced June 2024.

Comments: accepted to ACL2024 main

What Can Natural Language Processing Do for Peer Review?

Authors: Ilia Kuznetsov, Osama Mohammed Afzal, Koen Dercksen, Nils Dycke, Alexander Goldberg, Tom Hope, Dirk Hovy, Jonathan K. Kummerfeld, Anne Lauscher, Kevin Leyton-Brown, Sheng Lu, Mausam, Margot Mieskes, Aurélie Névéol, Danish Pruthi, Lizhen Qu, Roy Schwartz, Noah A. Smith, Thamar Solorio, **gyan Wang, Xiaodan Zhu, Anna Rogers, Nihar B. Shah, Iryna Gurevych

Abstract: The number of scientific articles produced every year is growing rapidly. Providing quality control over them is crucial for scientists and, ultimately, for the public good. In modern science, this process is largely delegated to peer review -- a distributed procedure in which each submission is evaluated by several independent experts in the field. Peer review is widely used, yet it is hard, time… ▽ More The number of scientific articles produced every year is growing rapidly. Providing quality control over them is crucial for scientists and, ultimately, for the public good. In modern science, this process is largely delegated to peer review -- a distributed procedure in which each submission is evaluated by several independent experts in the field. Peer review is widely used, yet it is hard, time-consuming, and prone to error. Since the artifacts involved in peer review -- manuscripts, reviews, discussions -- are largely text-based, Natural Language Processing has great potential to improve reviewing. As the emergence of large language models (LLMs) has enabled NLP assistance for many new tasks, the discussion on machine-assisted peer review is picking up the pace. Yet, where exactly is help needed, where can NLP help, and where should it stand aside? The goal of our paper is to provide a foundation for the future efforts in NLP for peer-reviewing assistance. We discuss peer review as a general process, exemplified by reviewing at AI conferences. We detail each step of the process from manuscript submission to camera-ready revision, and discuss the associated challenges and opportunities for NLP assistance, illustrated by existing work. We then turn to the big challenges in NLP for peer review as a whole, including data acquisition and licensing, operationalization and experimentation, and ethical issues. To help consolidate community efforts, we create a companion repository that aggregates key datasets pertaining to peer review. Finally, we issue a detailed call for action for the scientific community, NLP and AI researchers, policymakers, and funding bodies to help bring the research in NLP for peer review forward. We hope that our work will help set the agenda for research in machine-assisted scientific quality control in the age of AI, within the NLP community and beyond. △ Less

Submitted 10 May, 2024; originally announced May 2024.

Document Structure in Long Document Transformers

Authors: Jan Buchmann, Max Eichler, Jan-Micha Bodensohn, Ilia Kuznetsov, Iryna Gurevych

Abstract: Long documents often exhibit structure with hierarchically organized elements of different functions, such as section headers and paragraphs. Despite the omnipresence of document structure, its role in natural language processing (NLP) remains opaque. Do long-document Transformer models acquire an internal representation of document structure during pre-training? How can structural information be… ▽ More Long documents often exhibit structure with hierarchically organized elements of different functions, such as section headers and paragraphs. Despite the omnipresence of document structure, its role in natural language processing (NLP) remains opaque. Do long-document Transformer models acquire an internal representation of document structure during pre-training? How can structural information be communicated to a model after pre-training, and how does it influence downstream performance? To answer these questions, we develop a novel suite of probing tasks to assess structure-awareness of long-document Transformers, propose general-purpose structure infusion methods, and evaluate the effects of structure infusion on QASPER and Evidence Inference, two challenging long-document NLP tasks. Results on LED and LongT5 suggest that they acquire implicit understanding of document structure during pre-training, which can be further enhanced by structure infusion, leading to improved end-task performance. To foster research on the role of document structure in NLP modeling, we make our data and code publicly available. △ Less

Submitted 31 January, 2024; originally announced January 2024.

Comments: Accepted at EACL 2024. Code and data: http://github.com/UKPLab/eacl2024-doc-structure

Silicon Nanoantenna Mix Arrays for a Trifecta of Quantum Emitter Enhancements

Authors: Zhaogang Dong, Sergey Gorelik, Ramón Paniagua-Dominguez, Johnathan Yik, **fa Ho, Febiana Tjiptoharsono, Emmanuel Lassalle, Soroosh Daqiqeh Rezaei, Darren C. J. Neo, ** Bai, Arseniy I. Kuznetsov, Joel K. W. Yang

Abstract: Dielectric nanostructures have demonstrated optical antenna effects due to Mie resonances. Preliminary investigations on dielectric nanoantennas have been carried out for a trifecta of enhancements, i.e., simultaneous enhancements in absorption, emission directionality and radiative decay rates of quantum emitters. However, these investigations are limited by fragile substrates or low Purcell fact… ▽ More Dielectric nanostructures have demonstrated optical antenna effects due to Mie resonances. Preliminary investigations on dielectric nanoantennas have been carried out for a trifecta of enhancements, i.e., simultaneous enhancements in absorption, emission directionality and radiative decay rates of quantum emitters. However, these investigations are limited by fragile substrates or low Purcell factor, which is extremely important for exciting quantum emitters electrically. In this paper, we present a Si mix antenna array to achieve the trifecta enhancement of ~1200 fold with a Purcell factor of ~47. The antenna design incorporates ~10 nm gaps within which fluorescent molecules strongly absorb the pump laser energy through a resonant mode. In the emission process, the antenna array increases the radiative decay rates of the fluorescence molecules via Purcell effect and provides directional emission through a separate mode. This work could lead to novel CMOS compatible platforms for enhancing fluorescence for biological and chemical applications. △ Less

Submitted 10 October, 2023; originally announced October 2023.

Comments: 20 pages, 4 figures

Journal ref: Nano Letters 21, 4853-4860 (2021)

Large field-of-view and multi-color imaging with GaP quadratic metalenses

Authors: Anton V. Baranikov, Egor Khaidarov, Emmanuel Lassalle, Damien Eschimese, Joel Yeo, N. Duane Loh, Ramon Paniagua-Dominguez, Arseniy I. Kuznetsov

Abstract: Metalenses, in order to compete with conventional bulk optics in commercial imaging systems, often require large field of view (FOV) and broadband operation simultaneously. However, strong chromatic and coma aberrations present in common metalens designs have so far limited their widespread use. Stacking of metalenses as one of the possible solutions increases the overall complexity of the optical… ▽ More Metalenses, in order to compete with conventional bulk optics in commercial imaging systems, often require large field of view (FOV) and broadband operation simultaneously. However, strong chromatic and coma aberrations present in common metalens designs have so far limited their widespread use. Stacking of metalenses as one of the possible solutions increases the overall complexity of the optical system and hinders the main benefit of reduced thickness and light weight. To tackle both issues, here we propose a single-layer imaging system utilizing a recently developed class of metalenses providing large field of view. Using it, we demonstrate full-color imaging with a FOV of 100 degrees. This approach, empowered by computational imaging techniques, produce high quality images, both in terms of color reproduction and sharpness. Suitable for real-time unpolarized light operation with the standard color filters present in prevalent camera systems, our results might enable a pathway for consumer electronics applications of this emerging technology. △ Less

Submitted 26 May, 2023; originally announced May 2023.

Comments: 22 pages (main), 5 figures (main)

An efficient neural optimizer for resonant nanostructures: demonstration of highly-saturated red silicon structural color

Authors: Ronghui Lin, Vytautas Valuckas, Thi Thu Ha Do, Arash Nemati, Arseniy I. Kuznetsov, **ghua Teng, Son Tung Ha

Abstract: Freeform nanostructures have the potential to support complex resonances and their interactions, which are crucial for achieving desired spectral responses. However, the design optimization of such structures is nontrivial and computationally intensive. Furthermore, the current "black box" design approaches for freeform nanostructures often neglect the underlying physics. Here, we present a hybrid… ▽ More Freeform nanostructures have the potential to support complex resonances and their interactions, which are crucial for achieving desired spectral responses. However, the design optimization of such structures is nontrivial and computationally intensive. Furthermore, the current "black box" design approaches for freeform nanostructures often neglect the underlying physics. Here, we present a hybrid data-efficient neural optimizer for resonant nanostructures by combining a reinforcement learning algorithm and Powell's local optimization technique. As a case study, we design and experimentally demonstrate silicon nanostructures with a highly-saturated red color. Specifically, we achieved CIE color coordinates of (0.677, 0.304)-close to the ideal Schrodinger's red, with polarization independence, high reflectance (>85%), and a large viewing angle (i.e., up to ~ 25deg). The remarkable performance is attributed to underlying generalized multipolar interferences within each nanostructure rather than the collective array effects. Based on that, we were able to demonstrate pixel size down to ~400 nm, corresponding to a printing resolution of 65,000 pixels per inch. Moreover, the proposed design model requires only ~300 iterations to effectively search a 13-dimensional design space - an order of magnitude more efficient than the previously reported approaches. Our work significantly extends the free-form optical design toolbox for high-performance flat-optical components and metadevices. △ Less

Submitted 26 April, 2023; originally announced April 2023.

Electrically Tunable Reflective Metasurfaces with Continuous and Full Phase Modulation for High-efficiency Wavefront Control at Visible Frequencies

Authors: Parikshit Moitra, Xuewu Xu, Rasna Maruthiyodan Veetil, Xinan Liang, Tobias W. W. Mass, Arseniy I. Kuznetsov, Ramon Paniagua-Dominguez

Abstract: All-dielectric optical metasurfaces can locally control the amplitude and phase of light at the nanoscale, enabling arbitrary wavefront sha**. However, lack of post-fabrication tunability has limited the true potential of metasurfaces for many applications. Here, we utilize a thin liquid crystal (LC) layer as a tunable medium surrounding the metasurface to achieve a phase-only spatial light modu… ▽ More All-dielectric optical metasurfaces can locally control the amplitude and phase of light at the nanoscale, enabling arbitrary wavefront sha**. However, lack of post-fabrication tunability has limited the true potential of metasurfaces for many applications. Here, we utilize a thin liquid crystal (LC) layer as a tunable medium surrounding the metasurface to achieve a phase-only spatial light modulator (SLM) with high reflection in the visible frequency, exhibiting an active and continuous resonance tuning with associated 2pi phase control and uncoupled amplitude. Dynamic wavefront sha** is demonstrated by programming 96 individually addressable electrodes with a small pixel pitch of ~1 um. The small pixel size is facilitated by the reduced LC thickness, strongly suppressing crosstalk among pixels. This device is used to demonstrate dynamic beam steering with wide field-of-view and high absolute diffraction efficiencies. We believe that our demonstration may pave the way towards realizing next generation, high-resolution SLMs, with wide applications in dynamic holography, tunable optics and light detection and ranging (LiDAR), to mention a few. △ Less

Submitted 24 March, 2023; originally announced March 2023.

CARE: Collaborative AI-Assisted Reading Environment

Authors: Dennis Zyska, Nils Dycke, Jan Buchmann, Ilia Kuznetsov, Iryna Gurevych

Abstract: Recent years have seen impressive progress in AI-assisted writing, yet the developments in AI-assisted reading are lacking. We propose inline commentary as a natural vehicle for AI-based reading assistance, and present CARE: the first open integrated platform for the study of inline commentary and reading. CARE facilitates data collection for inline commentaries in a commonplace collaborative read… ▽ More Recent years have seen impressive progress in AI-assisted writing, yet the developments in AI-assisted reading are lacking. We propose inline commentary as a natural vehicle for AI-based reading assistance, and present CARE: the first open integrated platform for the study of inline commentary and reading. CARE facilitates data collection for inline commentaries in a commonplace collaborative reading environment, and provides a framework for enhancing reading with NLP-based assistance, such as text classification, generation or question answering. The extensible behavioral logging allows unique insights into the reading and commenting behavior, and flexible configuration makes the platform easy to deploy in new scenarios. To evaluate CARE in action, we apply the platform in a user study dedicated to scholarly peer review. CARE facilitates the data collection and study of inline commentary in NLP, extrinsic evaluation of NLP assistance, and application prototy**. We invite the community to explore and build upon the open source implementation of CARE. △ Less

Submitted 24 February, 2023; originally announced February 2023.

CiteBench: A benchmark for Scientific Citation Text Generation

Authors: Martin Funkquist, Ilia Kuznetsov, Yufang Hou, Iryna Gurevych

Abstract: Science progresses by building upon the prior body of knowledge documented in scientific publications. The acceleration of research makes it hard to stay up-to-date with the recent developments and to summarize the ever-growing body of prior work. To address this, the task of citation text generation aims to produce accurate textual summaries given a set of papers-to-cite and the citing paper cont… ▽ More Science progresses by building upon the prior body of knowledge documented in scientific publications. The acceleration of research makes it hard to stay up-to-date with the recent developments and to summarize the ever-growing body of prior work. To address this, the task of citation text generation aims to produce accurate textual summaries given a set of papers-to-cite and the citing paper context. Due to otherwise rare explicit anchoring of cited documents in the citing paper, citation text generation provides an excellent opportunity to study how humans aggregate and synthesize textual knowledge from sources. Yet, existing studies are based upon widely diverging task definitions, which makes it hard to study this task systematically. To address this challenge, we propose CiteBench: a benchmark for citation text generation that unifies multiple diverse datasets and enables standardized evaluation of citation text generation models across task designs and domains. Using the new benchmark, we investigate the performance of multiple strong baselines, test their transferability between the datasets, and deliver new insights into the task definition and evaluation to guide future research in citation text generation. We make the code for CiteBench publicly available at https://github.com/UKPLab/citebench. △ Less

Submitted 3 November, 2023; v1 submitted 19 December, 2022; originally announced December 2022.

Comments: To appear in EMNLP

Room-temperature strong coupling in a single photon emitter-dielectric metasurface system

Authors: T. Thu Ha Do, Milad Nonahal, Chi Li, Vytautas Valuckas, Arseniy I. Kuznetsov, Hai Son Nguyen, Igor Aharonovich, Son Tung Ha

Abstract: Single-photon sources with high brightness and long coherence time are promising qubit candidates for quantum technology. To this end, interfacing emitters with high-finesse cavities is required, especially in the strong coupling regime, which so far has only been limited to cryogenic temperatures. Here, we experimentally demonstrate, at room temperature, strong coupling between a single photon em… ▽ More Single-photon sources with high brightness and long coherence time are promising qubit candidates for quantum technology. To this end, interfacing emitters with high-finesse cavities is required, especially in the strong coupling regime, which so far has only been limited to cryogenic temperatures. Here, we experimentally demonstrate, at room temperature, strong coupling between a single photon emitter and a novel cavity based on optical bound states in the continuum. A remarkably large Rabi splitting of ~4 meV is achieved thanks to the combination of the narrow linewidth and large oscillator strength of emitters in hexagonal boron nitride and the efficient photon trap** of the cavity. Our findings unveil new opportunities to realise scalable quantum devices and explore fundamentally new regimes of strong coupling in quantum systems at room-temperature. △ Less

Submitted 13 December, 2022; originally announced December 2022.

NLPeer: A Unified Resource for the Computational Study of Peer Review

Authors: Nils Dycke, Ilia Kuznetsov, Iryna Gurevych

Abstract: Peer review constitutes a core component of scholarly publishing; yet it demands substantial expertise and training, and is susceptible to errors and biases. Various applications of NLP for peer reviewing assistance aim to support reviewers in this complex process, but the lack of clearly licensed datasets and multi-domain corpora prevent the systematic study of NLP for peer review. To remedy this… ▽ More Peer review constitutes a core component of scholarly publishing; yet it demands substantial expertise and training, and is susceptible to errors and biases. Various applications of NLP for peer reviewing assistance aim to support reviewers in this complex process, but the lack of clearly licensed datasets and multi-domain corpora prevent the systematic study of NLP for peer review. To remedy this, we introduce NLPeer -- the first ethically sourced multidomain corpus of more than 5k papers and 11k review reports from five different venues. In addition to the new datasets of paper drafts, camera-ready versions and peer reviews from the NLP community, we establish a unified data representation and augment previous peer review datasets to include parsed and structured paper representations, rich metadata and versioning information. We complement our resource with implementations and analysis of three reviewing assistance tasks, including a novel guided skimming task. Our work paves the path towards systematic, multi-faceted, evidence-based study of peer review in NLP and beyond. The data and code are publicly available. △ Less

Submitted 19 May, 2023; v1 submitted 12 November, 2022; originally announced November 2022.

Comments: Accepted at the ACL 2023 main conference

An Inclusive Notion of Text

Authors: Ilia Kuznetsov, Iryna Gurevych

Abstract: Natural language processing (NLP) researchers develop models of grammar, meaning and communication based on written text. Due to task and data differences, what is considered text can vary substantially across studies. A conceptual framework for systematically capturing these differences is lacking. We argue that clarity on the notion of text is crucial for reproducible and generalizable NLP. Towa… ▽ More Natural language processing (NLP) researchers develop models of grammar, meaning and communication based on written text. Due to task and data differences, what is considered text can vary substantially across studies. A conceptual framework for systematically capturing these differences is lacking. We argue that clarity on the notion of text is crucial for reproducible and generalizable NLP. Towards that goal, we propose common terminology to discuss the production and transformation of textual data, and introduce a two-tier taxonomy of linguistic and non-linguistic elements that are available in textual sources and can be used in NLP modeling. We apply this taxonomy to survey existing work that extends the notion of text beyond the conservative language-centered view. We outline key desiderata and challenges of the emerging inclusive approach to text in NLP, and suggest community-level reporting as a crucial next step to consolidate the discussion. △ Less

Submitted 17 May, 2023; v1 submitted 10 November, 2022; originally announced November 2022.

Comments: To appear in ACL 2023

Hybrid Dielectric-Plasmonic Nanoantenna with Multiresonances for Subwavelength Photon Sources

Authors: Pavel A. Dmitriev, Emmanuel Lassalle, Lu Ding, Zhenying Pan, Darren C. J. Neo, Vytautas Valuckas, Ramón Paniagua-Dominguez, Joel K. W. Yang, Hilmi Volkan Demir, Arseniy I. Kuznetsov

Abstract: The enhancement of the photoluminescence of quantum dots induced by an optical nanoantenna has been studied considerably, but there is still significant interest in optimizing and miniaturizing such structures, especially when accompanied by an experimental demonstration. Most of the realizations use plasmonic platforms, and some also use all-dielectric nanoantennas, but hybrid dielectric-plasmoni… ▽ More The enhancement of the photoluminescence of quantum dots induced by an optical nanoantenna has been studied considerably, but there is still significant interest in optimizing and miniaturizing such structures, especially when accompanied by an experimental demonstration. Most of the realizations use plasmonic platforms, and some also use all-dielectric nanoantennas, but hybrid dielectric-plasmonic (subwavelength) nanostructures have been very little explored. In this paper, we propose and demonstrate single subwavelength hybrid dielectric-plasmonic optical nanoantennas coupled to localized quantum dot emitters that constitute efficient and bright unidirectional photon sources under optical pum**. To achieve this, we devised a silicon nanoring sitting on a gold mirror with a 10 nm gap in-between, where an assembly of colloidal quantum dots is embedded. Such a structure supports both (radiative) antenna mode and (nonradiative) gap mode resonances, which we exploit for the dual purpose of out-coupling the light emitted by the quantum dots into the far-field with out-of-plane directivity, and for enhancing the excitation of the dots by the optical pump. Moreover, almost independent control of the resonance spectral positions can be achieved by simple tuning of geometrical parameters such as the ring inner and outer diameters, allowing us to conveniently adjust these resonances with respect to the quantum dots emission and absorption wavelengths. Using the proposed architecture, we obtain experimentally average fluorescence enhancement factors up to $654\times$ folds mainly due to high radiative efficiencies, and associated with a directional emission of the photoluminescence into a cone of $\pm 17\degree$ in the direction normal to the sample plane. We believe the solution presented here to be viable and relevant for the next generation of light-emitting devices. △ Less

Submitted 28 February, 2023; v1 submitted 27 June, 2022; originally announced June 2022.

Comments: 39 pages, 4 figures

Guided Exploration in Reinforcement Learning via Monte Carlo Critic Optimization

Authors: Igor Kuznetsov

Abstract: The class of deep deterministic off-policy algorithms is effectively applied to solve challenging continuous control problems. Current approaches commonly utilize random noise as an exploration method, which has several drawbacks, including the need for manual adjustment for a given task and the absence of exploratory calibration during the training process. We address these challenges by proposin… ▽ More The class of deep deterministic off-policy algorithms is effectively applied to solve challenging continuous control problems. Current approaches commonly utilize random noise as an exploration method, which has several drawbacks, including the need for manual adjustment for a given task and the absence of exploratory calibration during the training process. We address these challenges by proposing a novel guided exploration method that uses an ensemble of Monte Carlo Critics for calculating exploratory action correction. The proposed method enhances the traditional exploration scheme by dynamically adjusting exploration. Subsequently, we present a novel algorithm that leverages the proposed exploratory module for both policy and critic modification. The presented algorithm demonstrates superior performance compared to modern reinforcement learning algorithms across a variety of problems in the DMControl suite. △ Less

Submitted 6 May, 2024; v1 submitted 25 June, 2022; originally announced June 2022.

Comments: Short version of this paper is accepted to AAMAS 2024

Polarization-Tunable Perovskite Light-Emitting Metatransistor

Authors: Maciej Klein, Yutao Wang, **gyi Tian, Son Tung Ha, Ramón Paniagua-Domínguez, Arseniy I. Kuznetsov, Giorgio Adamo, Cesare Soci

Abstract: Emerging immersive visual communication technologies require light sources with complex functionality for dynamic control of polarization, directivity, wavefront, spectrum, and intensity of light. Currently, this is mostly achieved by free space bulk optic elements, limiting the adoption of these technologies. Flat optics based on artificially structured metasurfaces that operate at the sub-wavele… ▽ More Emerging immersive visual communication technologies require light sources with complex functionality for dynamic control of polarization, directivity, wavefront, spectrum, and intensity of light. Currently, this is mostly achieved by free space bulk optic elements, limiting the adoption of these technologies. Flat optics based on artificially structured metasurfaces that operate at the sub-wavelength scale are a viable solution, however their integration into electrically driven devices remains challenging. Here we demonstrate a radically new approach of monolithic integration of a dielectric metasurface into a perovskite light-emitting transistor. We show that nanogratings directly structured on top of the transistor channel yield an 8-fold increase of electroluminescence intensity and dynamic tunability of polarization. This new light-emitting metatransistor device concept opens unlimited opportunities for light management strategies based on metasurface design and integration. △ Less

Submitted 23 June, 2022; originally announced June 2022.

Tunable wavefront control in the visible spectrum using low-loss chalcogenide phase change metasurfaces

Authors: Parikshit Moitra, Yunzheng Wang, Xinan Liang, Li Lu, Alyssa Poh, Tobias W. W. Mass, Robert E. Simpson, Arseniy I. Kuznetsov, Ramon Paniagua-Dominguez

Abstract: All-dielectric metasurfaces provide unique solutions for advanced wavefront manipulation of light with complete control of amplitude and phase at sub-wavelength scales. One limitation, however, for most of these devices is the lack of any post-fabrication tunability of their response. To break this limit, a promising approach is employing phase-change-materials (PCM), which provide a fast, low ene… ▽ More All-dielectric metasurfaces provide unique solutions for advanced wavefront manipulation of light with complete control of amplitude and phase at sub-wavelength scales. One limitation, however, for most of these devices is the lack of any post-fabrication tunability of their response. To break this limit, a promising approach is employing phase-change-materials (PCM), which provide a fast, low energy and non-volatile means to endow metasurfaces with a switching mechanism. In this regard, great advancements have been done in the mid infrared and near infrared spectrum using different chalcogenides. In the visible spectral range, however, very few devices have demonstrated full phase manipulation, high efficiencies, and reversible switching. Here, we experimentally demonstrate a tunable all-dielectric Huygens metasurface made of antimony sulfide (Sb2S3) PCM, a low loss and high-index material in the visible spectral range with a large contrast (nearly 0.5) between its amorphous and crystalline states. We show close to 2pi phase modulation with high associated transmittance and use it to create switchable beam steering and holographic display devices. These novel chalcogenide PCM metasurfaces have the potential to emerge as a platform for next generation spatial light modulators and to impact application areas such as tunable and adaptive flat optics, LiDAR, and many more. △ Less

Submitted 15 June, 2022; originally announced June 2022.

Miniaturizing Color-Sensitive Photodetectors via Hybrid Nanoantennas towards Sub-micron Dimensions

Authors: **fa Ho, Zhaogang Dong, Hai Sheng Leong, Jun Zhang, Febiana Tjiptoharsono, Soroosh Daqiqeh Rezaei, Ken Choon Hwa Goh, Mengfei Wu, Shiqiang Li, **gyee Chee, Calvin Pei Yu Wong, Arseniy I. Kuznetsov, Joel K. W. Yang

Abstract: Digital camera sensors utilize color filters on photodiodes to achieve color selectivity. As color filters and photosensitive silicon layers are separate elements, these sensors suffer from optical cross-talk, which sets limits to the minimum pixel size. In this paper, we report hybrid silicon-aluminum nanostructures in the extreme limit of zero distance between color filters and sensors. This des… ▽ More Digital camera sensors utilize color filters on photodiodes to achieve color selectivity. As color filters and photosensitive silicon layers are separate elements, these sensors suffer from optical cross-talk, which sets limits to the minimum pixel size. In this paper, we report hybrid silicon-aluminum nanostructures in the extreme limit of zero distance between color filters and sensors. This design could essentially achieve sub micron pixel dimensions and minimize the optical cross-talk originated from tilt illuminations. The designed hybrid silicon-aluminum nanostructure has dual functionalities. Crucially, it supports a hybrid Mie-plasmon resonance of magnetic dipole to achieve the color-selective light absorption, generating electron hole pairs. Simultaneously, the silicon-aluminum interface forms a Schottky barrier for charge separation and photodetection. This design could potentially replace the traditional dye based filters for camera sensors at ultra-high pixel densities with advanced functionalities in sensing polarization and directionality, as well as UV selectivity via interband plasmons of silicon. △ Less

Submitted 7 June, 2022; originally announced June 2022.

Comments: 32 pages, 4 Figures (main text), 7 Figures (supplementary)

Revise and Resubmit: An Intertextual Model of Text-based Collaboration in Peer Review

Authors: Ilia Kuznetsov, Jan Buchmann, Max Eichler, Iryna Gurevych

Abstract: Peer review is a key component of the publishing process in most fields of science. The increasing submission rates put a strain on reviewing quality and efficiency, motivating the development of applications to support the reviewing and editorial work. While existing NLP studies focus on the analysis of individual texts, editorial assistance often requires modeling interactions between pairs of t… ▽ More Peer review is a key component of the publishing process in most fields of science. The increasing submission rates put a strain on reviewing quality and efficiency, motivating the development of applications to support the reviewing and editorial work. While existing NLP studies focus on the analysis of individual texts, editorial assistance often requires modeling interactions between pairs of texts -- yet general frameworks and datasets to support this scenario are missing. Relationships between texts are the core object of the intertextuality theory -- a family of approaches in literary studies not yet operationalized in NLP. Inspired by prior theoretical work, we propose the first intertextual model of text-based collaboration, which encompasses three major phenomena that make up a full iteration of the review-revise-and-resubmit cycle: pragmatic tagging, linking and long-document version alignment. While peer review is used across the fields of science and publication formats, existing datasets solely focus on conference-style review in computer science. Addressing this, we instantiate our proposed model in the first annotated multi-domain corpus in journal-style post-publication open peer review, and provide detailed insights into the practical aspects of intertextual annotation. Our resource is a major step towards multi-domain, fine-grained applications of NLP in editorial support for peer review, and our intertextual framework paves the path for general-purpose modeling of text-based collaboration. Our corpus and accompanying code are publicly available. △ Less

Submitted 31 May, 2022; v1 submitted 22 April, 2022; originally announced April 2022.

Yes-Yes-Yes: Proactive Data Collection for ACL Rolling Review and Beyond

Authors: Nils Dycke, Ilia Kuznetsov, Iryna Gurevych

Abstract: The shift towards publicly available text sources has enabled language processing at unprecedented scale, yet leaves under-serviced the domains where public and openly licensed data is scarce. Proactively collecting text data for research is a viable strategy to address this scarcity, but lacks systematic methodology taking into account the many ethical, legal and confidentiality-related aspects o… ▽ More The shift towards publicly available text sources has enabled language processing at unprecedented scale, yet leaves under-serviced the domains where public and openly licensed data is scarce. Proactively collecting text data for research is a viable strategy to address this scarcity, but lacks systematic methodology taking into account the many ethical, legal and confidentiality-related aspects of data collection. Our work presents a case study on proactive data collection in peer review -- a challenging and under-resourced NLP domain. We outline ethical and legal desiderata for proactive data collection and introduce "Yes-Yes-Yes", the first donation-based peer reviewing data collection workflow that meets these requirements. We report on the implementation of Yes-Yes-Yes at ACL Rolling Review and empirically study the implications of proactive data collection for the dataset size and the biases induced by the donation behavior on the peer reviewing platform. △ Less

Submitted 20 October, 2022; v1 submitted 27 January, 2022; originally announced January 2022.

Comments: Accepted at Findings of EMNLP 2022

High resolution multispectral spatial light modulators based on tunable Fabry-Perot nanocavities

Authors: Shampy Mansha, Parikshit Moitra, Xuewu Xu, Tobias W. W. Mass, Rasna Maruthiyodan Veetil, Xinan Liang, Shi-Qiang Li, Ramon Paniagua-Dominguez, Arseniy I. Kuznetsov

Abstract: Spatial light modulators (SLMs) are the most relevant technology for dynamic wavefront manipulation. They find diverse applications ranging from novel displays to optical and quantum communications. Among commercial SLMs for phase modulation, Liquid Crystal on Silicon (LCoS) offers the smallest pixel size and, thus, the most precise phase map** and largest field of view (FOV). Further pixel mini… ▽ More Spatial light modulators (SLMs) are the most relevant technology for dynamic wavefront manipulation. They find diverse applications ranging from novel displays to optical and quantum communications. Among commercial SLMs for phase modulation, Liquid Crystal on Silicon (LCoS) offers the smallest pixel size and, thus, the most precise phase map** and largest field of view (FOV). Further pixel miniaturization, however, is not possible in these devices due to inter-pixel cross-talks, which follow from the high driving voltages needed to modulate the thick liquid crystal (LC) cells that are necessary for full phase control. Newly introduced metasurface-based SLMs provide means for pixel miniaturization by modulating the phase via resonance tuning. These devices, however, are intrinsically monochromatic, limiting their use in applications requiring multi-wavelength operation. Here, we introduce a novel design allowing small pixel and multi-spectral operation. Based on LC-tunable Fabry-Perot nanocavities engineered to support multiple resonances across the visible range (including red, green and blue wavelengths), our design provides continuous 2pi phase modulation with high reflectance at each of the operating wavelengths. Experimentally, we realize a device with 96 pixels (~1μm pitch) that can be individually addressed by electrical biases. Using it, we first demonstrate multi-spectral programmable beam steering with FOV~18 degrees and absolute efficiencies exceeding 40%. Then, we reprogram the device to achieve multi-spectral lensing with tunable focal distance and efficiencies ~27%. Our design paves the way towards a new class of SLM for future applications in displays, optical computing and beyond. △ Less

Submitted 28 September, 2021; originally announced September 2021.

Comments: 43 pages, 15 figures

Bound state in the continuum in slab waveguide enables low-threshold quantum-dot lasing

Authors: Mengfei Wu, Lu Ding, Randy P. Sabatini, Laxmi Kishore Sagar, Golam Bappi, Ramon Paniagua-Dominguez, Edward H. Sargent, Arseniy I. Kuznetsov

Abstract: Colloidal quantum dots (CQDs) are a promising gain material for solution-processed, wavelength-tunable lasers, with potential application in displays, communications, and biomedical devices. In this work, we combine a CQD film with an array of titanium dioxide (TiO2) nanoantennas to achieve lasing via bound states in the continuum (BICs), which are symmetry-protected cavity modes with giant qualit… ▽ More Colloidal quantum dots (CQDs) are a promising gain material for solution-processed, wavelength-tunable lasers, with potential application in displays, communications, and biomedical devices. In this work, we combine a CQD film with an array of titanium dioxide (TiO2) nanoantennas to achieve lasing via bound states in the continuum (BICs), which are symmetry-protected cavity modes with giant quality factors. Here, the BICs arise from slab waveguide modes in the planar film, coupled to the periodic nanoantenna array. We engineer the thickness of the CQD film and size of the nanoantennas to achieve a BIC with good spatial and spectral overlap with the CQDs, based on a 2nd-order TE-polarized waveguide mode. We obtain room-temperature lasing with a low threshold of approximately 11 kW/cm2 (peak intensity) under 5 ns-pulsed optical excitation. This work sheds light on the optical modes in solution-processed, distributed-feedback lasers, and highlights BICs as effective, versatile, surface-emitting lasing modes. △ Less

Submitted 27 September, 2021; originally announced September 2021.

Comments: 30 pages, 7 figures

Journal ref: Nano Letters, 2021

Assisting Decision Making in Scholarly Peer Review: A Preference Learning Perspective

Authors: Nils Dycke, Edwin Simpson, Ilia Kuznetsov, Iryna Gurevych

Abstract: Peer review is the primary means of quality control in academia; as an outcome of a peer review process, program and area chairs make acceptance decisions for each paper based on the review reports and scores they received. Quality of scientific work is multi-faceted; coupled with the subjectivity of reviewing, this makes final decision making difficult and time-consuming. To support this final st… ▽ More Peer review is the primary means of quality control in academia; as an outcome of a peer review process, program and area chairs make acceptance decisions for each paper based on the review reports and scores they received. Quality of scientific work is multi-faceted; coupled with the subjectivity of reviewing, this makes final decision making difficult and time-consuming. To support this final step of peer review, we formalize it as a paper ranking problem. We introduce a novel, multi-faceted generic evaluation framework for ranking submissions based on peer reviews that takes into account effectiveness, efficiency and fairness. We propose a preference learning perspective on the task that considers both review texts and scores to alleviate the inevitable bias and noise in reviews. Our experiments on peer review data from the ACL 2018 conference demonstrate the superiority of our preference-learning-based approach over baselines and prior work, while highlighting the importance of using both review texts and scores to rank submissions. △ Less

Submitted 27 May, 2022; v1 submitted 2 September, 2021; originally announced September 2021.

Sorrel: an IDE Plugin for Managing Licenses and Detecting License Incompatibilities

Authors: Dmitry Pogrebnoy, Ivan Kuznetsov, Yaroslav Golubev, Vladislav Tankov, Timofey Bryksin

Abstract: Software development is a complex process that includes many different tasks besides just writing code. One of the aspects of software engineering is selecting and managing licenses for the given project. In this paper, we present Sorrel - a plugin for managing licenses and detecting potential incompatibilities for IntelliJ IDEA, a popular Java IDE. The plugin scans the project in search of inform… ▽ More Software development is a complex process that includes many different tasks besides just writing code. One of the aspects of software engineering is selecting and managing licenses for the given project. In this paper, we present Sorrel - a plugin for managing licenses and detecting potential incompatibilities for IntelliJ IDEA, a popular Java IDE. The plugin scans the project in search of information about the project license and the licenses of its libraries. If the project does not yet have a license, the plugin provides the developer with recommendations for choosing the most suitable open license, and if there is a license, it informs the programmer about potential licensing violations. The tool makes it easier for developers to choose a proper license for a project and avoid most of the licensing errors - all inside the familiar IDE editor. The plugin and its source code are available online on GitHub: https://github.com/JetBrains-Research/sorrel. A demonstration video can be found at https://youtu.be/doUeAwPjcPE. △ Less

Submitted 28 July, 2021; originally announced July 2021.

Comments: 5 pages, 3 figures

80-degree field-of-view transmissive metasurface-based spatial light modulator

Authors: Anton V. Baranikov, Shi-Qiang Li, Damien Eschimese, Xuewu Xu, Simon Thiele, Simon Ristok, Rasna Maruthiyodan Veetil, Tobias W. W. Mass, Parikshit Moitra, Harald Giessen, Ramon Paniagua-Dominguez, Arseniy I. Kuznetsov

Abstract: Compact, lightweight and high-performance spatial light modulators (SLMs) are crucial for modern optical technologies. The drive for pixel miniaturization, necessary to improve their performance, has led to a promising alternative, active optical metasurfaces, which enable tunable subwavelength wavefront manipulation. Here, we demonstrate an all-solid-state programmable transmissive SLM device bas… ▽ More Compact, lightweight and high-performance spatial light modulators (SLMs) are crucial for modern optical technologies. The drive for pixel miniaturization, necessary to improve their performance, has led to a promising alternative, active optical metasurfaces, which enable tunable subwavelength wavefront manipulation. Here, we demonstrate an all-solid-state programmable transmissive SLM device based on Huygens dielectric metasurfaces. The metasurface features electrical tunability, provided by mature liquid crystals (LCs) technology. In contrast to conventional LC SLMs, our device enables high resolution with a pixel size of ~1 um. We demonstrate its performance by realizing programmable beam steering, which exhibits high side mode suppression ratio of ~6 dB. By complementing the device with a 3D printed doublet microlens, fabricated using two-photon polymerization, we enhance the field of view up to ~80 degrees. The developed prototype paves the way to compact, efficient and multifunctional devices for next generation augmented reality displays, light detection and ranging (LiDAR) systems and optical computing. △ Less

Submitted 23 July, 2021; originally announced July 2021.

Comments: 25 pages, 8 figures

Phase-change perovskite tunable microlaser

Authors: **gyi Tian, Giorgio Adamo, Hailong Liu, Mengfei Wu, Maciej Klein, Jie Deng, Norman Soo Seng Ang, Ramón Paniagua-Domínguez, Hong Liu, Arseniy I. Kuznetsov, Cesare Soci

Abstract: Since the invention of the laser, adoption of new gain media and device architectures has provided solutions to a variety of applications requiring specific power, size, spectral, spatial, and temporal tunability. Here we introduce a fundamentally new type of tunable semiconductor laser based on a phase-change perovskite metasurface that acts simultaneously as gain medium and optical cavity. As a… ▽ More Since the invention of the laser, adoption of new gain media and device architectures has provided solutions to a variety of applications requiring specific power, size, spectral, spatial, and temporal tunability. Here we introduce a fundamentally new type of tunable semiconductor laser based on a phase-change perovskite metasurface that acts simultaneously as gain medium and optical cavity. As a proof of principle demonstration, we fabricate a subwavelength-thin perovskite metasurface supporting bound states in the continuum (BICs). Upon the perovskite structural phase transitions, both its refractive index and gain vary substantially, inducing fast (1.35 nm/K rate) and broad spectral tunability (>15 nm in the near-infrared), deterministic spatial mode hop** between polarization vortexes, and hysteretic optical bistability of the microlaser. These features highlight the uniqueness of phase-change perovskite tunable lasers, which may find wide applications in compact and low-cost optical multiplexers, sensors, memories, and LIDARs. △ Less

Submitted 12 July, 2021; originally announced July 2021.

Solving Continuous Control with Episodic Memory

Authors: Igor Kuznetsov, Andrey Filchenkov

Abstract: Episodic memory lets reinforcement learning algorithms remember and exploit promising experience from the past to improve agent performance. Previous works on memory mechanisms show benefits of using episodic-based data structures for discrete action problems in terms of sample-efficiency. The application of episodic memory for continuous control with a large action space is not trivial. Our study… ▽ More Episodic memory lets reinforcement learning algorithms remember and exploit promising experience from the past to improve agent performance. Previous works on memory mechanisms show benefits of using episodic-based data structures for discrete action problems in terms of sample-efficiency. The application of episodic memory for continuous control with a large action space is not trivial. Our study aims to answer the question: can episodic memory be used to improve agent's performance in continuous control? Our proposed algorithm combines episodic memory with Actor-Critic architecture by modifying critic's objective. We further improve performance by introducing episodic-based replay buffer prioritization. We evaluate our algorithm on OpenAI gym domains and show greater sample-efficiency compared with the state-of-the art model-free off-policy algorithms. △ Less

Submitted 16 June, 2021; originally announced June 2021.

Comments: To appear in the 30th International Joint Conference on Artificial Intelligence (IJCAI 2021)

Photonic Bound-States-in-the-Continuum Observed with an Electron Nanoprobe

Authors: Zhaogang Dong, Zackaria Mahfoud, Ramón Paniagua-Domínguez, Hongtao Wang, Antonio I. Fernández-Domínguez, Sergey Gorelik, Son Tung Ha, Febiana Tjiptoharsono, Arseniy I. Kuznetsov, Michel Bosman, Joel K. W. Yang

Abstract: Bound-states-in-the-continuum (BIC)is a wave-mechanical concept that generates resonances with vanishing spectral linewidths. It has many practical applications in Optics, such as narrow-band filters, mirror-less lasing, and nonlinear harmonic generation. As true BIC optical modes non-radiative and confined to the near field of nanostructures, they cannot be excited using propagating light. As a r… ▽ More Bound-states-in-the-continuum (BIC)is a wave-mechanical concept that generates resonances with vanishing spectral linewidths. It has many practical applications in Optics, such as narrow-band filters, mirror-less lasing, and nonlinear harmonic generation. As true BIC optical modes non-radiative and confined to the near field of nanostructures, they cannot be excited using propagating light. As a result, their direct experimental observation has been elusive. Rather than using light, we demonstrate probing BIC modes on arrays of silicon nanoantennas using a focused beam of electrons in a tranmission electron microscope. By combining cathodoluminescence (CL) and monochromated electron energy-loss spectroscopy (EELS) with controlled nanofabrication, we provide direct experimental evidence of "true" BIC modes, and demonstrate a BIC mode in the visible spectrum at 720 nm. The ability to observe and quantify these guided resonances with a spatial precision more than two orders of magnitude higher than previous far-field measurements allows the probing of individual elements in the nano-antenna arrays. The high-resolution experimental results are supported by numerical simulations as well as multipolar decomposition analysis, allowing us to demonstrate that the coherent interaction length of the quasi-BIC resonance requires at least 6 neighboring antenna elements, achieving over 60 times higher emissivity than for unpatterned silicon. △ Less

Submitted 10 May, 2021; originally announced May 2021.

Comments: 30 pages, 4 figures in the main text, 6 figures in the supplementary information

Effect of transport current on suppression of superconductivity with ultrashort laser pulse

Authors: P F Kartsev, I O Kuznetsov

Abstract: We study the suppression of superconductivity with ultrashort laser pulse in the presence of transport current. The theoretical model is based on the BCS relations for the superconducting state coupled with kinetic equations for nonequilibrium Bogoliubov quasiparticles and phonons. The results of numerical simulation for picosecond and femtosecond laser pulses of optical and infrared ranges are gi… ▽ More We study the suppression of superconductivity with ultrashort laser pulse in the presence of transport current. The theoretical model is based on the BCS relations for the superconducting state coupled with kinetic equations for nonequilibrium Bogoliubov quasiparticles and phonons. The results of numerical simulation for picosecond and femtosecond laser pulses of optical and infrared ranges are given. We discuss the effects of main problem parameters, including the current density. △ Less

Submitted 15 March, 2021; originally announced March 2021.

Comments: 14 pages, 6 figures

Imaging properties of large field-of-view quadratic metalenses and their applications to fingerprint detection

Authors: Emmanuel Lassalle, Tobias W. W. Mass, Damien Eschimese, Anton V. Baranikov, Egor Khaidarov, Shiqiang Li, Ramon Paniagua-Dominguez, Arseniy I. Kuznetsov

Abstract: Dielectric metasurfaces, extremely thin nanostructured dielectric surfaces, hold promise to replace conventional refractive optics, such as lenses, due to their high performance and compactness. However, designing large field-of-view (FOV) metalenses, which are of particular importance when imaging relatively big objects at short distances, remains one of the most critical challenges. Recently, me… ▽ More Dielectric metasurfaces, extremely thin nanostructured dielectric surfaces, hold promise to replace conventional refractive optics, such as lenses, due to their high performance and compactness. However, designing large field-of-view (FOV) metalenses, which are of particular importance when imaging relatively big objects at short distances, remains one of the most critical challenges. Recently, metalenses implementing a quadratic phase profile have been put forward to solve this problem with a single element, but despite their theoretical ability to provide $180^\circ\,$FOV, imaging over very large FOV has not been demonstrated yet. In this work, we provide an in-depth analysis of the imaging properties of quadratic metalenses and, in particular, show that due to their intrinsic barrel distortion or fish-eye effect, there is a fundamental trade-off between the FOV achievable in a given imaging configuration and the optical resolution of the metalens and/or the detector resolution. To illustrate how to harness the full potential of quadratic metalenses, we apply these considerations to the fingerprint detection problem, and demonstrate experimentally the full imaging of a $5\,$mm fingerprint with features of the order of $100\,μ$m, with a metalens ten times smaller in size and located at a distance of only $2.5\,$mm away from the object. This constitutes the most compact system reported so far for the fingerprint detection. △ Less

Submitted 31 May, 2021; v1 submitted 16 February, 2021; originally announced February 2021.

Comments: 6 figures

Continuous wave second harmonic generation enabled by quasi-bound-states in the continuum on gallium phosphide metasurfaces

Authors: Aravind P. Anthur, Haizhong Zhang, Ramon Paniagua-Dominguez, Dmitry Kalashnikov, Son Tung Ha, Tobias Wilhelm Wolfgang Mass, Arseniy I. Kuznetsov, Leonid Krivitsky

Abstract: Resonant metasurfaces are an attractive platform for enhancing the non-linear optical processes, such as second harmonic generation (SHG), since they can generate very large local electromagnetic fields while relaxing the phase-matching requirements. Here, we take this platform a step closer to the practical applications by demonstrating visible range, continuous wave (CW) SHG. We do so by combini… ▽ More Resonant metasurfaces are an attractive platform for enhancing the non-linear optical processes, such as second harmonic generation (SHG), since they can generate very large local electromagnetic fields while relaxing the phase-matching requirements. Here, we take this platform a step closer to the practical applications by demonstrating visible range, continuous wave (CW) SHG. We do so by combining the attractive material properties of gallium phosphide with engineered, high quality-factor photonic modes enabled by bound states in the continuum. For the optimum case, we obtain efficiencies around 5e-5 % W$^{-1}$ when the system is pumped at 1200 nm wavelength with CW intensities of 1 kW/cm$^2$. Moreover, we measure external efficiencies as high as 0.1 % W$^{-1}$ with pump intensities of only 10 MW/cm$^2$ for pulsed irradiation. This efficiency is higher than the values previously reported for dielectric metasurfaces, but achieved here with pump intensities that are two orders of magnitude lower. △ Less

Submitted 3 February, 2021; originally announced February 2021.

Method of numerical simulation of interacting quantum gas kinetics on finite momentum lattice

Authors: I. O. Kuznetsov, P. F. Kartsev

Abstract: We present the efficient and universal numerical method for simulation of interacting quantum gas kinetics on a finite momentum lattice, based on the Boltzmann equation for occupation numbers. Usually, the study of models with two-particle interaction generates the excessive amount of terms in the equations essentially limiting the possible system size. Here we employ the original analytical trans… ▽ More We present the efficient and universal numerical method for simulation of interacting quantum gas kinetics on a finite momentum lattice, based on the Boltzmann equation for occupation numbers. Usually, the study of models with two-particle interaction generates the excessive amount of terms in the equations essentially limiting the possible system size. Here we employ the original analytical transformation to decrease the scaling index of the amount of calculations. As a result, lattice sizes as large as 48x48x48 can be simulated, allowing to study realistic problems with complex interaction models. The method was applied to the simulation of weakly interacting Fermi and Bose gases where we calculated the relaxation times depending on the momentum and temperature. △ Less

Submitted 28 November, 2020; originally announced November 2020.

Comments: 10 pages, 7 figures

Journal ref: Revised and extended version published as Phys. Rev. E 107, 055304 (2023)

Large-scale Huygens metasurfaces for holographic 3D near-eye displays

Authors: Weitao Song, Xinan Liang, Shiqiang Li, Dongdong Li, Ramon Paniagua-Dominguez, Keng Heng Lai, Qunying Lin, Yuan** Zheng, Arseniy I. Kuznetsov

Abstract: Novel display technologies aim at providing the users with increasingly immersive experiences. In this regard, it is a long-sought dream to generate three-dimensional (3D) scenes with high resolution and continuous depth, which can be overlaid with the real world. Current attempts to do so, however, fail in providing either truly 3D information, or a large viewing area and angle, strongly limiting… ▽ More Novel display technologies aim at providing the users with increasingly immersive experiences. In this regard, it is a long-sought dream to generate three-dimensional (3D) scenes with high resolution and continuous depth, which can be overlaid with the real world. Current attempts to do so, however, fail in providing either truly 3D information, or a large viewing area and angle, strongly limiting the user immersion. Here, we report a proof-of-concept solution for this problem, and realize a compact holographic 3D near-eye display with a large exit pupil of 10mm x 8.66mm. The 3D image is generated from a highly transparent Huygens metasurface hologram with large (>10^8) pixel count and subwavelength pixels, fabricated via deep-ultraviolet immersion photolithography on 300 mm glass wafers. We experimentally demonstrate high quality virtual 3D scenes with ~50k active data points and continuous depth ranging from 0.5m to 2m, overlaid with the real world and easily viewed by naked eye. To do so, we introduce a new design method for holographic near-eye displays that, inherently, is able to provide both parallax and accommodation cues, fundamentally solving the vergence-accommodation conflict that exists in current commercial 3D displays. △ Less

Submitted 9 October, 2020; originally announced October 2020.

Comments: 21 pages, 9 figures

Journal ref: Laser & Photonics Reviews, 2021

Generation of even and odd high harmonics in resonant metasurfaces using single and multiple ultra-intense laser pulses

Authors: Maxim R. Shcherbakov, Haizhong Zhang, Michael Tripepi, Giovanni Sartorello, Noah Talisa, Abdallah AlShafey, Zhiyuan Fan, Justin Twardowski, Leonid A. Krivitsky, Arseniy I. Kuznetsov, Enam Chowdhury, Gennady Shvets

Abstract: High harmonic generation (HHG) opens a window on the fundamental science of strong-field light-mater interaction and serves as a key building block for attosecond optics and metrology. Resonantly enhanced HHG from hot spots in nanostructures is an attractive route to overcoming the well-known limitations of gases and bulk solids. We demonstrate a nanoscale platform for highly efficient HHG driven… ▽ More High harmonic generation (HHG) opens a window on the fundamental science of strong-field light-mater interaction and serves as a key building block for attosecond optics and metrology. Resonantly enhanced HHG from hot spots in nanostructures is an attractive route to overcoming the well-known limitations of gases and bulk solids. We demonstrate a nanoscale platform for highly efficient HHG driven by strong mid-infrared laser pulses: an ultra-thin resonant gallium phosphide (GaP) metasurface. The wide bandgap and the lack of inversion symmetry of the GaP crystal enable the generation of even and odd harmonics covering a wide range of photon energies between 1.3 and 3 eV with minimal reabsorption. The resonantly enhanced conversion efficiency facilitates single-shot measurements that avoid material damage and pave the way to controllable transition between perturbative and non-perturbative regimes of light-matter interactions at the nanoscale. △ Less

Submitted 13 October, 2020; v1 submitted 8 August, 2020; originally announced August 2020.

Comments: 36 pages, 9 figures, 2 tables

Journal ref: Nature Communications 12, 4185 (2021)

Nonlinear interferometry with infrared metasurfaces

Authors: Anna V. Paterova, Dmitry A. Kalashnikov, Egor Khaidarov, Hongzhi Yang, Tobias W. W. Mass, Ramon Paniagua-Dominguez, Arseniy I. Kuznetsov, Leonid A. Krivitsky

Abstract: The optical elements comprised of sub-diffractive light scatterers, or metasurfaces, hold a promise to reduce the footprint and unfold new functionalities of optical devices. A particular interest is focused on metasurfaces for manipulation of phase and amplitude of light beams. Characterisation of metasurfaces can be performed using interferometry, which, however, may be cumbersome, specifically… ▽ More The optical elements comprised of sub-diffractive light scatterers, or metasurfaces, hold a promise to reduce the footprint and unfold new functionalities of optical devices. A particular interest is focused on metasurfaces for manipulation of phase and amplitude of light beams. Characterisation of metasurfaces can be performed using interferometry, which, however, may be cumbersome, specifically in the infrared (IR) range. Here, we realise a new method for characterising IR metasurfaces based on nonlinear interference, which uses accessible components for visible light. Correlated IR and visible photons are launched into a nonlinear interferometer so that the phase profile, imposed by the metasurface on the IR photons, modifies the interference at the visible photon wavelength. Furthermore, we show that this concept can be used for broadband manipulation of the intensity profile of a visible beam using a single IR metasurface. Our method unfolds the potential of quantum interferometry for the characterization of advanced optical elements. △ Less

Submitted 28 July, 2020; originally announced July 2020.

Comments: Submitted for publication

Modeling tau transport in the axon initial segment

Authors: Ivan A. Kuznetsov, Andrey V. Kuznetsov

Abstract: By assuming that tau protein can be in seven kinetic states, we developed a model of tau protein transport in the axon and in the axon initial segment (AIS). Two separate sets of kinetic constants were determined, one in the axon and the other in the AIS. This was done by fitting the model predictions in the axon with experimental results and by fitting the model predictions in the AIS with the as… ▽ More By assuming that tau protein can be in seven kinetic states, we developed a model of tau protein transport in the axon and in the axon initial segment (AIS). Two separate sets of kinetic constants were determined, one in the axon and the other in the AIS. This was done by fitting the model predictions in the axon with experimental results and by fitting the model predictions in the AIS with the assumed linear increase of the total tau concentration in the AIS. The calibrated model was used to make predictions about tau transport in the axon and in the AIS. To the best of our knowledge, this is the first paper that presents a mathematical model of tau transport in the AIS. Our modeling results suggest that binding of free tau to MTs creates a negative gradient of free tau in the AIS. This leads to diffusion-driven tau transport from the soma into the AIS. The model further suggests that slow axonal transport and diffusion-driven transport of tau work together in the AIS, moving tau anterogradely. Our numerical results predict an interplay between these two mechanisms: as the distance from the soma increases, the diffusion-driven transport decreases, while motor-driven transport becomes larger. Thus, the machinery in the AIS works as a pump, moving tau into the axon. △ Less

Submitted 1 October, 2020; v1 submitted 29 June, 2020; originally announced June 2020.

Comments: final accepted version

Journal ref: Mathematical Biosciences, vol. 329, article # 108468, 2020

Terahertz signatures of ultrafast Dirac fermion relaxation at the surface of topological insulators at room temperature

Authors: S. Kovalev, K. -J. Tielrooij, J. -C. Deinert, I. Ilyakov, N. Awari, M. Chen, A. Ponomaryov, M. Bawatna, T. V. A. G. de Oliveira, L. M. Eng, K. A. Kuznetsov, G. Kh. Kitaeva, P. I. Kuznetsov, H. A. Hafez, D. Turchinovich, M. Gensch

Abstract: Topologically-protected surface states present rich physics and promising spintronic, optoelectronic and photonic applications that require a proper understanding of their ultrafast carrier dynamics. Here, we investigate these dynamics in topological insulators (TIs) of the bismuth and antimony chalcogenide family, where we isolate the response of Dirac fermions at the surface from the response of… ▽ More Topologically-protected surface states present rich physics and promising spintronic, optoelectronic and photonic applications that require a proper understanding of their ultrafast carrier dynamics. Here, we investigate these dynamics in topological insulators (TIs) of the bismuth and antimony chalcogenide family, where we isolate the response of Dirac fermions at the surface from the response of bulk carriers by combining photoexcitation with below-bandgap terahertz (THz) photons with TI samples with varying Fermi level, including one sample with the Fermi level located within the bandgap. We identify distinctly faster relaxation of charge carriers in the topologically-protected Dirac surface states (few hundred femtoseconds), compared to bulk carriers (few picoseconds). In agreement with such fast cooling dynamics, we observe THz harmonic generation without any saturation effects for increasing incident fields, unlike graphene which exhibits strong saturation. This opens up promising avenues for increased THz nonlinear conversion efficiencies, and high-bandwidth optoelectronic and spintronic information and communication applications. △ Less

Submitted 1 March, 2021; v1 submitted 6 June, 2020; originally announced June 2020.

A Matter of Framing: The Impact of Linguistic Formalism on Probing Results

Authors: Ilia Kuznetsov, Iryna Gurevych

Abstract: Deep pre-trained contextualized encoders like BERT (Delvin et al., 2019) demonstrate remarkable performance on a range of downstream tasks. A recent line of research in probing investigates the linguistic knowledge implicitly learned by these models during pre-training. While most work in probing operates on the task level, linguistic tasks are rarely uniform and can be represented in a variety of… ▽ More Deep pre-trained contextualized encoders like BERT (Delvin et al., 2019) demonstrate remarkable performance on a range of downstream tasks. A recent line of research in probing investigates the linguistic knowledge implicitly learned by these models during pre-training. While most work in probing operates on the task level, linguistic tasks are rarely uniform and can be represented in a variety of formalisms. Any linguistics-based probing study thereby inevitably commits to the formalism used to annotate the underlying data. Can the choice of formalism affect probing results? To investigate, we conduct an in-depth cross-formalism layer probing study in role semantics. We find linguistically meaningful differences in the encoding of semantic role- and proto-role information by BERT depending on the formalism and demonstrate that layer probing can detect subtle differences between the implementations of the same linguistic formalism. Our results suggest that linguistic formalism is an important dimension in probing studies, along with the commonly used cross-task and cross-lingual experimental settings. △ Less

Submitted 30 April, 2020; originally announced April 2020.

Lasing action in single subwavelength particles supporting supercavity modes

Authors: Vasilii Mylnikov, Son Tung Ha, Zhenying Pan, Vytautas Valuckas, Ramon Paniagua-Dominguez, Hilmi Volkan Demir, Arseniy I. Kuznetsov

Abstract: On-chip light sources are critical for the realization of fully integrated photonic circuitry. So far, semiconductor miniaturized lasers have been mainly limited to sizes on the order of a few microns. Further reduction of sizes is challenging fundamentally due to the associated radiative losses. While using plasmonic metals helps to reduce radiative losses and sizes, they also introduce Ohmic los… ▽ More On-chip light sources are critical for the realization of fully integrated photonic circuitry. So far, semiconductor miniaturized lasers have been mainly limited to sizes on the order of a few microns. Further reduction of sizes is challenging fundamentally due to the associated radiative losses. While using plasmonic metals helps to reduce radiative losses and sizes, they also introduce Ohmic losses hindering real improvements. In this work, we show that, making use of quasi-bound states in the continuum, or supercavity modes, we circumvent these fundamental issues and realize the smallest purely semiconductor nanolaser thus far. Here, the nanolaser structure is based on a single semiconductor nanocylinder that intentionally takes advantage of the destructive interference between two supported optical modes, namely Fabry-Perot and Mie modes, to obtain a significant enhancement in the quality factor of the cavity. We experimentally demonstrate the concept and obtain optically pumped lasing action using GaAs at cryogenic temperatures. The optimal nanocylinder size is as small as 500 nm in diameter and only 330 nm in height with a lasing wavelength around 825 nm, corresponding to a size-to-wavelength ratio around 0.6. The obtained results pave the way for the development of smaller on-chip light sources free of Ohmic losses, which may find applications in the future photonic circuits. △ Less

Submitted 9 March, 2020; originally announced March 2020.

Comments: 29 pages, 9 figures

Journal ref: ACS Nano (2020)

Demonstration of second harmonic generationin gallium phosphide nano-waveguides

Authors: Aravind P. Anthur, Haizhong Zhang, Yuriy Akimov, Junrong Ong, Dmitry Kalashnikov, Arseniy I. Kuznetsov, Leonid Krivitsky

Abstract: We designed, fabricated and tested gallium phosphide (GaP) nano-waveguides for second harmonic generation (SHG). We demonstrate SHG in the visible range around 655 nm using low power continuous-wave pump in the optical communication O-band. Our structures utilize modal phase matching, such that lower order eigenmodes of the pump are phase matched to higher order eigenmodes of the second harmonic.… ▽ More We designed, fabricated and tested gallium phosphide (GaP) nano-waveguides for second harmonic generation (SHG). We demonstrate SHG in the visible range around 655 nm using low power continuous-wave pump in the optical communication O-band. Our structures utilize modal phase matching, such that lower order eigenmodes of the pump are phase matched to higher order eigenmodes of the second harmonic. We observe phase matched SHG for different combinations of interacting modes by varying the widths of the waveguides and tuning the wavelength of the pump. The presented results contribute to the development of integrated photonic platforms with efficient nonlinear wave-mixing processes for classical and quantum applications. △ Less

Submitted 22 June, 2020; v1 submitted 16 January, 2020; originally announced January 2020.

Journal ref: Opt. Express Vol. 29, Issue 7, pp. 10307-10320 (2021)

How old are dense core vesicles residing in en passant boutons: Simulation of the mean age of dense core vesicles in axonal arbors accounting for resident and transiting vesicle populations

Authors: I. A. Kuznetsov, A. V. Kuznetsov

Abstract: In neurons, neuropeptides are synthesized in the soma and are then transported along the axon in dense core vesicles (DCVs). DCVs are captured in varicosities located along the axon terminal called en passant boutons, which are active terminal sites that accumulate and release neurotransmitters. Recently developed experimental techniques allow for the estimation of the age of DCVs in various locat… ▽ More In neurons, neuropeptides are synthesized in the soma and are then transported along the axon in dense core vesicles (DCVs). DCVs are captured in varicosities located along the axon terminal called en passant boutons, which are active terminal sites that accumulate and release neurotransmitters. Recently developed experimental techniques allow for the estimation of the age of DCVs in various locations in the axon terminal. Accurate simulation of the mean age of DCVs in boutons requires the development of a model that would account for resident, transiting-anterograde, and transiting-retrograde DCV populations. In this paper, such a model is developed. The model is applied to simulating DCV transport in Drosophila type II motoneurons. The model simulates DCV transport and capture in the axon terminals and makes it possible to predict the age density distribution of DCVs in en passant boutons as well as DCV's mean age in boutons. The predicted prevalence of older organelles in distal boutons may explain the "dying back" pattern of axonal degeneration observed in dopaminergic neurons in Parkinson's disease. The predicted difference of two hours between the age of older DCVs residing in distal boutons and the age of younger DCVs residing in proximal boutons is consistent with an approximate estimate of age difference deduced from experimental observations. The age density of resident DCVs is found to be bimodal, which is because DCVs are captured from two transiting states: the anterograde transiting state that contains younger DCVs and the retrograde transiting state that contains older DCVs. △ Less

Submitted 16 October, 2020; v1 submitted 29 December, 2019; originally announced December 2019.

Comments: Final, accepted manuscript version

Journal ref: Proc. R. Soc. A, vol. 476: 20200454, 2020

Global deformations of a Lie algebra of type $\bar{A_5}$

Authors: N. G. Chebochko, M. I. Kuznetsov

Abstract: It is shown that the orbits of the space of local deformations of the Lie algebra $\bar{A_5}$ over an algebraically closed field $K$ of characteristic 2 with respect to the automorphism group $\mathrm{PGL} (6)$ correspond to $\mathrm{GL} (V)$-orbits of tri-vectors of a 6-dimensional space. For local deformations corresponding to tri-vectors of rank $ρ<6$, integrability is proved and global deforma… ▽ More It is shown that the orbits of the space of local deformations of the Lie algebra $\bar{A_5}$ over an algebraically closed field $K$ of characteristic 2 with respect to the automorphism group $\mathrm{PGL} (6)$ correspond to $\mathrm{GL} (V)$-orbits of tri-vectors of a 6-dimensional space. For local deformations corresponding to tri-vectors of rank $ρ<6$, integrability is proved and global deformations are constructed. △ Less

Submitted 5 January, 2020; v1 submitted 25 December, 2019; originally announced December 2019.

Voltage plateaus on V(I) curves of long quasi-one-dimensional superconducting wires (without microwave irradiation)

Authors: V. I. Kuznetsov, A. A. Firsov

Abstract: Almost constant voltage plateaus on the V(I) curves of long quasi-one-dimensional superconducting aluminum wires placed in magnetic field at temperatures T slightly below the critical superconducting temperature Tc were found which were unexpected for the sample geometry and parameters of the experiment. The plateaus are assumed to be subharmonics of a superconducting gap and arise due to the mult… ▽ More Almost constant voltage plateaus on the V(I) curves of long quasi-one-dimensional superconducting aluminum wires placed in magnetic field at temperatures T slightly below the critical superconducting temperature Tc were found which were unexpected for the sample geometry and parameters of the experiment. The plateaus are assumed to be subharmonics of a superconducting gap and arise due to the multiple Andreev reflection and strong quasiparticle overheating in the wire nonequilibrium region. The plateaus are evidence of coexistence of superconductivity and dissipation in such wires. The results presented in the paper could not be described by existing theories. △ Less

Submitted 12 August, 2019; originally announced August 2019.

Comments: 7 pages, 4 figures; the author's translation of the original Russian text: V.I. Kuznetsov, A.A. Firsov, Pis'ma v Zhurnal Eksperimental'noi i Teoreticheskoi Fiziki 104, 721 (2016)

Journal ref: JETP Lett. 104, 709 (2016) (original Russian text: Pis'ma v Zhurnal Eksperimental'noi i Teoreticheskoi Fiziki 104, 721 (2016))

Quantum magnetoresistive (hc/2e)/m periodic oscillations in a superconducting ring

Authors: V. I. Kuznetsov, O. V. Trofimov

Abstract: It was experimentally found that quantum magnetoresistive hc/2e periodic oscillations of the Little-Parks type in a superconducting mesoscopic ring with decreasing temperature and increasing applied dc current are modified to the sum of harmonic (hc/2e)/m periodic oscillations. Multiple Andreev reflection can be a possible cause of this effect. It was experimentally found that quantum magnetoresistive hc/2e periodic oscillations of the Little-Parks type in a superconducting mesoscopic ring with decreasing temperature and increasing applied dc current are modified to the sum of harmonic (hc/2e)/m periodic oscillations. Multiple Andreev reflection can be a possible cause of this effect. △ Less

Submitted 12 August, 2019; originally announced August 2019.

Comments: 8 pages, 8 figures; the author's translation of the original Russian text: V.I. Kuznetsov, O.V. Trofimov, Pis'ma v Zhurnal Eksperimental'noi i Teoreticheskoi Fiziki 110, 47 (2019)

Journal ref: JETP Lett. 110, 54 (2019) (original Russian text: Pis'ma v Zhurnal Eksperimental'noi i Teoreticheskoi Fiziki 110, 47 (2019))

Control of LED Emission with Functional Dielectric Metasurfaces

Authors: Egor Khaidarov, Zhengtong Liu, Ramon Paniagua-Dominguez, Son Tung Ha, Vytautas Valuckas, Xinan Liang, Yuriy Akimov, ** Bai, Ching Eng Png, Hilmi Volkan Demir, Arseniy I. Kuznetsov

Abstract: The improvement of light-emitting diodes (LEDs) is one of the major goals of optoelectronics and photonics research. While emission rate enhancement is certainly one of the targets, in this regard, for LED integration to complex photonic devices, one would require to have, additionally, precise control of the wavefront of the emitted light. Metasurfaces are spatial arrangements of engineered scatt… ▽ More The improvement of light-emitting diodes (LEDs) is one of the major goals of optoelectronics and photonics research. While emission rate enhancement is certainly one of the targets, in this regard, for LED integration to complex photonic devices, one would require to have, additionally, precise control of the wavefront of the emitted light. Metasurfaces are spatial arrangements of engineered scatters that may enable this light manipulation capability with unprecedented resolution. Most of these devices, however, are only able to function properly under irradiation of light with a large spatial coherence, typically normally incident lasers. LEDs, on the other hand, have angularly broad, Lambertian-like emission patterns characterized by a low spatial coherence, which makes the integration of metasurface devices on LED architectures extremely challenging. A novel concept for metasurface integration on LED is proposed, using a cavity to increase the LED spatial coherence through an angular collimation. Due to the resonant character of the cavity, extending the spatial coherence of the emitted light does not come at the price of any reduction in the total emitted power. The experimental demonstration of the proposed concept is implemented on a GaP LED architecture including a hybrid metallic-Bragg cavity. By integrating a silicon metasurface on top we demonstrate two different functionalities of these compact devices: directional LED emission at a desired angle and LED emission of a vortex beam with an orbital angular momentum. The presented concept is general, being applicable to other incoherent light sources and enabling metasurfaces designed for plane waves to work with incoherent light emitters. △ Less

Submitted 18 July, 2019; originally announced July 2019.

Comments: 29 pages, 6 figures

Singular limits of the quasi-linear Kolmogorov-type equation with a source term

Authors: Ivan Kuznetsov, Sergey Sazhenkov

Abstract: Existence, uniqueness and stability of kinetic and entropy solutions to the boundary value problem for the Kolmogorov-type genuinely nonlinear ultra-parabolic equation with a smooth source term is established. After this, we consider the case when the source term contains a small positive parameter and collapses to the Dirac delta-function, as this parameter tends to zero. In this case, the limiti… ▽ More Existence, uniqueness and stability of kinetic and entropy solutions to the boundary value problem for the Kolmogorov-type genuinely nonlinear ultra-parabolic equation with a smooth source term is established. After this, we consider the case when the source term contains a small positive parameter and collapses to the Dirac delta-function, as this parameter tends to zero. In this case, the limiting passage from the original equation with the smooth source to the impulsive ultra-parabolic equation is fulfilled and rigorously justified. The proofs rely on the method of kinetic equation and on the compensated compactness techniques for genuinely nonlinear equations. △ Less

Submitted 9 July, 2019; originally announced July 2019.

MSC Class: 35D30; 35K70; 35R12

Conditioning of Reinforcement Learning Agents and its Policy Regularization Application

Authors: Arip Asadulaev, Igor Kuznetsov, Gideon Stein, Andrey Filchenkov

Abstract: The outcome of Jacobian singular values regularization was studied for supervised learning problems. It also was shown that Jacobian conditioning regularization can help to avoid the ``mode-collapse'' problem in Generative Adversarial Networks. In this paper, we try to answer the following question: Can information about policy conditioning help to shape a more stable and general policy of reinfor… ▽ More The outcome of Jacobian singular values regularization was studied for supervised learning problems. It also was shown that Jacobian conditioning regularization can help to avoid the ``mode-collapse'' problem in Generative Adversarial Networks. In this paper, we try to answer the following question: Can information about policy conditioning help to shape a more stable and general policy of reinforcement learning agents? To answer this question, we conduct a study of Jacobian conditioning behavior during policy optimization. To the best of our knowledge, this is the first work that research condition number in reinforcement learning agents. We propose a conditioning regularization algorithm and test its performance on the range of continuous control tasks. Finally, we compare algorithms on the CoinRun environment with separated train end test levels to analyze how conditioning regularization contributes to agents' generalization. △ Less

Submitted 13 July, 2020; v1 submitted 12 June, 2019; originally announced June 2019.

Interpretable Few-Shot Learning via Linear Distillation

Authors: Arip Asadulaev, Igor Kuznetsov, Andrey Filchenkov

Abstract: It is important to develop mathematically tractable models than can interpret knowledge extracted from the data and provide reasonable predictions. In this paper, we present a Linear Distillation Learning, a simple remedy to improve the performance of linear neural networks. Our approach is based on using a linear function for each class in a dataset, which is trained to simulate the output of a t… ▽ More It is important to develop mathematically tractable models than can interpret knowledge extracted from the data and provide reasonable predictions. In this paper, we present a Linear Distillation Learning, a simple remedy to improve the performance of linear neural networks. Our approach is based on using a linear function for each class in a dataset, which is trained to simulate the output of a teacher linear network for each class separately. We tested our model on MNIST and Omniglot datasets in the Few-Shot learning manner. It showed better results than other interpretable models such as classical Logistic Regression. △ Less

Submitted 10 October, 2019; v1 submitted 12 June, 2019; originally announced June 2019.

First measurement of near-threshold J/$ψ$ exclusive photoproduction off the proton

Authors: The GlueX Collaboration, A. Ali, M. Amaryan, E. G. Anassontzis, A. Austregesilo, M. Baalouch, F. Barbosa, J. Barlow, A. Barnes, E. Barriga, T. D. Beattie, V. V. Berdnikov, T. Black, W. Boeglin, M. Boer, W. J. Briscoe, T. Britton, W. K. Brooks, B. E. Cannon, N. Cao, E. Chudakov, S. Cole, O. Cortes, V. Crede, M. M. Dalton , et al. (110 additional authors not shown)

Abstract: We report on the measurement of the $γp \rightarrow J/ψp$ cross section from $E_γ= 11.8$ GeV down to the threshold at $8.2$ GeV using a tagged photon beam with the GlueX experiment. We find the total cross section falls toward the threshold less steeply than expected from two-gluon exchange models. The differential cross section $dσ/dt$ has an exponential slope of $1.67 \pm 0.39$ GeV$^{-2}$ at… ▽ More We report on the measurement of the $γp \rightarrow J/ψp$ cross section from $E_γ= 11.8$ GeV down to the threshold at $8.2$ GeV using a tagged photon beam with the GlueX experiment. We find the total cross section falls toward the threshold less steeply than expected from two-gluon exchange models. The differential cross section $dσ/dt$ has an exponential slope of $1.67 \pm 0.39$ GeV$^{-2}$ at $10.7$ GeV average energy. The LHCb pentaquark candidates $P_c^+$ can be produced in the $s$-channel of this reaction. We see no evidence for them and set model-dependent upper limits on their branching fractions $\mathcal{B}(P_c^+ \rightarrow J/ψp)$ and cross sections $σ(γp \to P_c^+)\times\mathcal{B}(P_c^+ \to J/ψp) $. △ Less

Submitted 10 September, 2019; v1 submitted 26 May, 2019; originally announced May 2019.

Comments: 12 pages, 7 figures

Journal ref: Phys. Rev. Lett. 123, 072001 (2019)