Showing 1–50 of 69 results for author: Stevens, S

Reducibility points and characteristic $p$ local fields I- Simple supercuspidal representations of symplectic groups

Authors: Corinne Blondel, Guy Henniart, Shaun Stevens

Abstract: Let $F$ be a non-Archimedean local field with odd characteristic $p$. Let $N$ be a positive integer and $G=Sp_{2N}(F)$. By work of Lomelí on $γ$-factors of pairs and converse theorems, a generic supercuspidal representation $π$ of $G$ has a transfer to a smooth irreducible representation $Π_π$ of $GL_{2N+1}(F)$. In turn the Weil-Deligne representation $Σ_π$ associated to $Π_π$ by the Langlands cor… ▽ More Let $F$ be a non-Archimedean local field with odd characteristic $p$. Let $N$ be a positive integer and $G=Sp_{2N}(F)$. By work of Lomelí on $γ$-factors of pairs and converse theorems, a generic supercuspidal representation $π$ of $G$ has a transfer to a smooth irreducible representation $Π_π$ of $GL_{2N+1}(F)$. In turn the Weil-Deligne representation $Σ_π$ associated to $Π_π$ by the Langlands correspondence determines a Langlands parameter $φ_π$ for $π$. That process produces a Langlands correspondence for generic cuspidal representations of $G$. In this paper we take $π$ to be simple in the sense of Gross and Reeder, and from the explicit construction of $π$ we describe $Π_π$ explicitly. The method we use is the same as in our previous paper arXiv:2310.20455, where we treated the case where $F$ is a $p$-adic field, and $π$ a simple supercuspidal representation of $G=Sp_{2N}(F)$. It relies on a criterion due to Moeglin on the reducibility of representations parabolically induced from $GL_M(F)\times G$ for varying positive integers $M$. We extend this criterion to the case when $F$ has any positive characteristic. The main new feature consists in relating reducibility to $γ$-factors for pairs. △ Less

Submitted 22 June, 2024; originally announced June 2024.

Comments: 8 pages

MSC Class: 22E50; 11F70

Active Diffusion Subsampling

Authors: Oisin Nolan, Tristan S. W. Stevens, Wessel L. van Nierop, Ruud J. G. van Sloun

Abstract: Subsampling is commonly used to mitigate costs associated with data acquisition, such as time or energy requirements, motivating the development of algorithms for estimating the fully-sampled signal of interest $x$ from partially observed measurements $y$. In maximum-entropy sampling, one selects measurement locations that are expected to have the highest entropy, so as to minimize uncertainty abo… ▽ More Subsampling is commonly used to mitigate costs associated with data acquisition, such as time or energy requirements, motivating the development of algorithms for estimating the fully-sampled signal of interest $x$ from partially observed measurements $y$. In maximum-entropy sampling, one selects measurement locations that are expected to have the highest entropy, so as to minimize uncertainty about $x$. This approach relies on an accurate model of the posterior distribution over future measurements, given the measurements observed so far. Recently, diffusion models have been shown to produce high-quality posterior samples of high-dimensional signals using guided diffusion. In this work, we propose Active Diffusion Subsampling (ADS), a method for performing active subsampling using guided diffusion in which the model tracks a distribution of beliefs over the true state of $x$ throughout the reverse diffusion process, progressively decreasing its uncertainty by choosing to acquire measurements with maximum expected entropy, and ultimately generating the posterior distribution $p(x | y)$. ADS can be applied using pre-trained diffusion models for any subsampling rate, and does not require task-specific retraining - just the specification of a measurement model. Furthermore, the maximum entropy sampling policy employed by ADS is interpretable, enhancing transparency relative to existing methods using black-box policies. Experimentally, we show that ADS outperforms fixed sampling strategies, and study an application of ADS in Magnetic Resonance Imaging acceleration using the fastMRI dataset, finding that ADS performs competitively with supervised methods. Code available at https://active-diffusion-subsampling.github.io/. △ Less

Submitted 20 June, 2024; originally announced June 2024.

Comments: 17 pages, 12 figures, preprint

Block decompositions for $p$-adic classical groups and their inner forms

Authors: David Helm, Robert Kurinczuk, Daniel Skodlerack, Shaun Stevens

Abstract: For an inner form $\mathrm{G}$ of a general linear group or classical group over a non-archimedean local field of odd residue characteristic, we decompose the category of smooth representations on $\mathbb{Z}[μ_{p^{\infty}},1/p]$-modules by endo-parameter. We prove that parabolic induction preserves these decompositions, and hence that it preserves endo-parameters. Moreover, we show that the decom… ▽ More For an inner form $\mathrm{G}$ of a general linear group or classical group over a non-archimedean local field of odd residue characteristic, we decompose the category of smooth representations on $\mathbb{Z}[μ_{p^{\infty}},1/p]$-modules by endo-parameter. We prove that parabolic induction preserves these decompositions, and hence that it preserves endo-parameters. Moreover, we show that the decomposition by endo-parameter is the $\overline{\mathbb{Z}}[1/p]$-block decomposition; and, for $\mathrm{R}$ an integral domain, introduce a graph whose connected components parameterize the $\mathrm{R}$-blocks, in particular including the cases $\mathrm{R}=\overline{\mathbb{Z}}_{\ell}$ and $\mathrm{R}=\overline{\mathbb{F}}_\ell$ for $\ell\neq p$. From our description, we deduce that the $\overline{\mathbb{Z}_\ell}$-blocks and $\overline{\mathbb{F}_\ell}$-blocks of $\mathrm{G}$ are in natural bijection, as had long been expected. Our methods also apply to the trivial endo-parameter (i.e., the depth zero subcategory) of any connected reductive $p$-adic group, providing an alternative approach to results of Dat and Lanard in depth zero. Finally, under a technical assumption (known for inner forms of general linear groups) we reduce the $\mathrm{R}$-block decomposition of $\mathrm{G}$ to depth zero. △ Less

Submitted 22 May, 2024; originally announced May 2024.

Comments: 54 pages

The cool and the cruel: separating hard parts of LWE secrets

Authors: Niklas Nolte, Mohamed Malhou, Emily Wenger, Samuel Stevens, Cathy Li, François Charton, Kristin Lauter

Abstract: Sparse binary LWE secrets are under consideration for standardization for Homomorphic Encryption and its applications to private computation. Known attacks on sparse binary LWE secrets include the sparse dual attack and the hybrid sparse dual-meet in the middle attack which requires significant memory. In this paper, we provide a new statistical attack with low memory requirement. The attack relie… ▽ More Sparse binary LWE secrets are under consideration for standardization for Homomorphic Encryption and its applications to private computation. Known attacks on sparse binary LWE secrets include the sparse dual attack and the hybrid sparse dual-meet in the middle attack which requires significant memory. In this paper, we provide a new statistical attack with low memory requirement. The attack relies on some initial lattice reduction. The key observation is that, after lattice reduction is applied to the rows of a q-ary-like embedded random matrix $\mathbf A$, the entries with high variance are concentrated in the early columns of the extracted matrix. This allows us to separate out the "hard part" of the LWE secret. We can first solve the sub-problem of finding the "cruel" bits of the secret in the early columns, and then find the remaining "cool" bits in linear time. We use statistical techniques to distinguish distributions to identify both the cruel and the cool bits of the secret. We provide concrete attack timings for recovering secrets in dimensions $n=256$, $512$, and $768$. For the lattice reduction stage, we leverage recent improvements in lattice reduction (e.g. flatter) applied in parallel. We also apply our new attack in the RLWE setting for $2$-power cyclotomic rings, showing that these RLWE instances are much more vulnerable to this attack than LWE. △ Less

Submitted 15 March, 2024; originally announced March 2024.

Salsa Fresca: Angular Embeddings and Pre-Training for ML Attacks on Learning With Errors

Authors: Samuel Stevens, Emily Wenger, Cathy Li, Niklas Nolte, Eshika Saxena, François Charton, Kristin Lauter

Abstract: Learning with Errors (LWE) is a hard math problem underlying recently standardized post-quantum cryptography (PQC) systems for key exchange and digital signatures. Prior work proposed new machine learning (ML)-based attacks on LWE problems with small, sparse secrets, but these attacks require millions of LWE samples to train on and take days to recover secrets. We propose three key methods -- bett… ▽ More Learning with Errors (LWE) is a hard math problem underlying recently standardized post-quantum cryptography (PQC) systems for key exchange and digital signatures. Prior work proposed new machine learning (ML)-based attacks on LWE problems with small, sparse secrets, but these attacks require millions of LWE samples to train on and take days to recover secrets. We propose three key methods -- better preprocessing, angular embeddings and model pre-training -- to improve these attacks, speeding up preprocessing by $25\times$ and improving model sample efficiency by $10\times$. We demonstrate for the first time that pre-training improves and reduces the cost of ML attacks on LWE. Our architecture improvements enable scaling to larger-dimension LWE problems: this work is the first instance of ML attacks recovering sparse binary secrets in dimension $n=1024$, the smallest dimension used in practice for homomorphic encryption applications of LWE where sparse binary secrets are proposed. △ Less

Submitted 1 February, 2024; originally announced February 2024.

Comments: 8 pages (main text)

BioCLIP: A Vision Foundation Model for the Tree of Life

Authors: Samuel Stevens, Jiaman Wu, Matthew J Thompson, Elizabeth G Campolongo, Chan Hee Song, David Edward Carlyn, Li Dong, Wasila M Dahdul, Charles Stewart, Tanya Berger-Wolf, Wei-Lun Chao, Yu Su

Abstract: Images of the natural world, collected by a variety of cameras, from drones to individual phones, are increasingly abundant sources of biological information. There is an explosion of computational methods and tools, particularly computer vision, for extracting biologically relevant information from images for science and conservation. Yet most of these are bespoke approaches designed for a specif… ▽ More Images of the natural world, collected by a variety of cameras, from drones to individual phones, are increasingly abundant sources of biological information. There is an explosion of computational methods and tools, particularly computer vision, for extracting biologically relevant information from images for science and conservation. Yet most of these are bespoke approaches designed for a specific task and are not easily adaptable or extendable to new questions, contexts, and datasets. A vision model for general organismal biology questions on images is of timely need. To approach this, we curate and release TreeOfLife-10M, the largest and most diverse ML-ready dataset of biology images. We then develop BioCLIP, a foundation model for the tree of life, leveraging the unique properties of biology captured by TreeOfLife-10M, namely the abundance and variety of images of plants, animals, and fungi, together with the availability of rich structured biological knowledge. We rigorously benchmark our approach on diverse fine-grained biology classification tasks and find that BioCLIP consistently and substantially outperforms existing baselines (by 16% to 17% absolute). Intrinsic evaluation reveals that BioCLIP has learned a hierarchical representation conforming to the tree of life, shedding light on its strong generalizability. https://imageomics.github.io/bioclip has models, data and code. △ Less

Submitted 14 May, 2024; v1 submitted 30 November, 2023; originally announced November 2023.

Comments: CVPR 2024 (oral) camera-ready version; data released

MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI

Authors: Xiang Yue, Yuansheng Ni, Kai Zhang, Tianyu Zheng, Ruoqi Liu, Ge Zhang, Samuel Stevens, Dongfu Jiang, Weiming Ren, Yuxuan Sun, Cong Wei, Botao Yu, Ruibin Yuan, Renliang Sun, Ming Yin, Boyuan Zheng, Zhenzhu Yang, Yibo Liu, Wenhao Huang, Huan Sun, Yu Su, Wenhu Chen

Abstract: We introduce MMMU: a new benchmark designed to evaluate multimodal models on massive multi-discipline tasks demanding college-level subject knowledge and deliberate reasoning. MMMU includes 11.5K meticulously collected multimodal questions from college exams, quizzes, and textbooks, covering six core disciplines: Art & Design, Business, Science, Health & Medicine, Humanities & Social Science, and… ▽ More We introduce MMMU: a new benchmark designed to evaluate multimodal models on massive multi-discipline tasks demanding college-level subject knowledge and deliberate reasoning. MMMU includes 11.5K meticulously collected multimodal questions from college exams, quizzes, and textbooks, covering six core disciplines: Art & Design, Business, Science, Health & Medicine, Humanities & Social Science, and Tech & Engineering. These questions span 30 subjects and 183 subfields, comprising 30 highly heterogeneous image types, such as charts, diagrams, maps, tables, music sheets, and chemical structures. Unlike existing benchmarks, MMMU focuses on advanced perception and reasoning with domain-specific knowledge, challenging models to perform tasks akin to those faced by experts. The evaluation of 14 open-source LMMs as well as the proprietary GPT-4V(ision) and Gemini highlights the substantial challenges posed by MMMU. Even the advanced GPT-4V and Gemini Ultra only achieve accuracies of 56% and 59% respectively, indicating significant room for improvement. We believe MMMU will stimulate the community to build next-generation multimodal foundation models towards expert artificial general intelligence. △ Less

Submitted 13 June, 2024; v1 submitted 27 November, 2023; originally announced November 2023.

Comments: CVPR 2024 Oral

A Simple Interpretable Transformer for Fine-Grained Image Classification and Analysis

Authors: Dipanjyoti Paul, Arpita Chowdhury, Xinqi Xiong, Feng-Ju Chang, David Carlyn, Samuel Stevens, Kaiya L. Provost, Anuj Karpatne, Bryan Carstens, Daniel Rubenstein, Charles Stewart, Tanya Berger-Wolf, Yu Su, Wei-Lun Chao

Abstract: We present a novel usage of Transformers to make image classification interpretable. Unlike mainstream classifiers that wait until the last fully connected layer to incorporate class information to make predictions, we investigate a proactive approach, asking each class to search for itself in an image. We realize this idea via a Transformer encoder-decoder inspired by DEtection TRansformer (DETR)… ▽ More We present a novel usage of Transformers to make image classification interpretable. Unlike mainstream classifiers that wait until the last fully connected layer to incorporate class information to make predictions, we investigate a proactive approach, asking each class to search for itself in an image. We realize this idea via a Transformer encoder-decoder inspired by DEtection TRansformer (DETR). We learn "class-specific" queries (one for each class) as input to the decoder, enabling each class to localize its patterns in an image via cross-attention. We name our approach INterpretable TRansformer (INTR), which is fairly easy to implement and exhibits several compelling properties. We show that INTR intrinsically encourages each class to attend distinctively; the cross-attention weights thus provide a faithful interpretation of the prediction. Interestingly, via "multi-head" cross-attention, INTR could identify different "attributes" of a class, making it particularly suitable for fine-grained classification and analysis, which we demonstrate on eight datasets. Our code and pre-trained models are publicly accessible at the Imageomics Institute GitHub site: https://github.com/Imageomics/INTR. △ Less

Submitted 14 June, 2024; v1 submitted 7 November, 2023; originally announced November 2023.

Comments: Accepted to International Conference on Learning Representations 2024 (ICLR 2024)

Simple cuspidal representations of symplectic groups: Langlands parameter

Authors: Corinne Blondel, Guy Henniart, Shaun Stevens

Abstract: Let $F$ be a non-archimedean local field of odd residual characteristic. We compute the Jordan set of a simple cuspidal representation of a symplectic group over $F$, using explicit computations of generators of the Hecke algebras of covers reflecting the parabolic induction under study. When $F$ is a $p$-adic field we obtain the Langlands parameter of the representation. Let $F$ be a non-archimedean local field of odd residual characteristic. We compute the Jordan set of a simple cuspidal representation of a symplectic group over $F$, using explicit computations of generators of the Hecke algebras of covers reflecting the parabolic induction under study. When $F$ is a $p$-adic field we obtain the Langlands parameter of the representation. △ Less

Submitted 31 October, 2023; originally announced October 2023.

Comments: 33 pages

MSC Class: 22E50; 11F70

Roll Up Your Sleeves: Working with a Collaborative and Engaging Task-Oriented Dialogue System

Authors: Lingbo Mo, Shijie Chen, Ziru Chen, Xiang Deng, Ashley Lewis, Sunit Singh, Samuel Stevens, Chang-You Tai, Zhen Wang, Xiang Yue, Tianshu Zhang, Yu Su, Huan Sun

Abstract: We introduce TacoBot, a user-centered task-oriented digital assistant designed to guide users through complex real-world tasks with multiple steps. Covering a wide range of cooking and how-to tasks, we aim to deliver a collaborative and engaging dialogue experience. Equipped with language understanding, dialogue management, and response generation components supported by a robust search engine, Ta… ▽ More We introduce TacoBot, a user-centered task-oriented digital assistant designed to guide users through complex real-world tasks with multiple steps. Covering a wide range of cooking and how-to tasks, we aim to deliver a collaborative and engaging dialogue experience. Equipped with language understanding, dialogue management, and response generation components supported by a robust search engine, TacoBot ensures efficient task assistance. To enhance the dialogue experience, we explore a series of data augmentation strategies using LLMs to train advanced neural models continuously. TacoBot builds upon our successful participation in the inaugural Alexa Prize TaskBot Challenge, where our team secured third place among ten competing teams. We offer TacoBot as an open-source framework that serves as a practical example for deploying task-oriented dialogue systems. △ Less

Submitted 29 July, 2023; originally announced July 2023.

Dehazing Ultrasound using Diffusion Models

Authors: Tristan S. W. Stevens, Faik C. Meral, Jason Yu, Iason Z. Apostolakis, Jean-Luc Robert, Ruud J. G. van Sloun

Abstract: Echocardiography has been a prominent tool for the diagnosis of cardiac disease. However, these diagnoses can be heavily impeded by poor image quality. Acoustic clutter emerges due to multipath reflections imposed by layers of skin, subcutaneous fat, and intercostal muscle between the transducer and heart. As a result, haze and other noise artifacts pose a real challenge to cardiac ultrasound imag… ▽ More Echocardiography has been a prominent tool for the diagnosis of cardiac disease. However, these diagnoses can be heavily impeded by poor image quality. Acoustic clutter emerges due to multipath reflections imposed by layers of skin, subcutaneous fat, and intercostal muscle between the transducer and heart. As a result, haze and other noise artifacts pose a real challenge to cardiac ultrasound imaging. In many cases, especially with difficult-to-image patients such as patients with obesity, a diagnosis from B-Mode ultrasound imaging is effectively rendered unusable, forcing sonographers to resort to contrast-enhanced ultrasound examinations or refer patients to other imaging modalities. Tissue harmonic imaging has been a popular approach to combat haze, but in severe cases is still heavily impacted by haze. Alternatively, denoising algorithms are typically unable to remove highly structured and correlated noise, such as haze. It remains a challenge to accurately describe the statistical properties of structured haze, and develop an inference method to subsequently remove it. Diffusion models have emerged as powerful generative models and have shown their effectiveness in a variety of inverse problems. In this work, we present a joint posterior sampling framework that combines two separate diffusion models to model the distribution of both clean ultrasound and haze in an unsupervised manner. Furthermore, we demonstrate techniques for effectively training diffusion models on radio-frequency ultrasound data and highlight the advantages over image data. Experiments on both \emph{in-vitro} and \emph{in-vivo} cardiac datasets show that the proposed dehazing method effectively removes haze while preserving signals from weakly reflected tissue. △ Less

Submitted 10 December, 2023; v1 submitted 20 July, 2023; originally announced July 2023.

Comments: 12pages, 15 figures, preprint IEEE submission

Mind2Web: Towards a Generalist Agent for the Web

Authors: Xiang Deng, Yu Gu, Boyuan Zheng, Shijie Chen, Samuel Stevens, Boshi Wang, Huan Sun, Yu Su

Abstract: We introduce Mind2Web, the first dataset for develo** and evaluating generalist agents for the web that can follow language instructions to complete complex tasks on any website. Existing datasets for web agents either use simulated websites or only cover a limited set of websites and tasks, thus not suitable for generalist web agents. With over 2,000 open-ended tasks collected from 137 websites… ▽ More We introduce Mind2Web, the first dataset for develo** and evaluating generalist agents for the web that can follow language instructions to complete complex tasks on any website. Existing datasets for web agents either use simulated websites or only cover a limited set of websites and tasks, thus not suitable for generalist web agents. With over 2,000 open-ended tasks collected from 137 websites spanning 31 domains and crowdsourced action sequences for the tasks, Mind2Web provides three necessary ingredients for building generalist web agents: 1) diverse domains, websites, and tasks, 2) use of real-world websites instead of simulated and simplified ones, and 3) a broad spectrum of user interaction patterns. Based on Mind2Web, we conduct an initial exploration of using large language models (LLMs) for building generalist web agents. While the raw HTML of real-world websites are often too large to be fed to LLMs, we show that first filtering it with a small LM significantly improves the effectiveness and efficiency of LLMs. Our solution demonstrates a decent level of performance, even on websites or entire domains the model has never seen before, but there is still a substantial room to improve towards truly generalizable agents. We open-source our dataset, model implementation, and trained models (https://osu-nlp-group.github.io/Mind2Web) to facilitate further research on building a generalist agent for the web. △ Less

Submitted 9 December, 2023; v1 submitted 9 June, 2023; originally announced June 2023.

Comments: Website: https://osu-nlp-group.github.io/Mind2Web. Updated with supplementary material. NeurIPS'23 Spotlight

Memorization for Good: Encryption with Autoregressive Language Models

Authors: Samuel Stevens, Yu Su

Abstract: Over-parameterized neural language models (LMs) can memorize and recite long sequences of training data. While such memorization is normally associated with undesired properties such as overfitting and information leaking, our work casts memorization as an unexplored capability of LMs. We propose the first symmetric encryption algorithm with autoregressive language models (SELM). We show that auto… ▽ More Over-parameterized neural language models (LMs) can memorize and recite long sequences of training data. While such memorization is normally associated with undesired properties such as overfitting and information leaking, our work casts memorization as an unexplored capability of LMs. We propose the first symmetric encryption algorithm with autoregressive language models (SELM). We show that autoregressive LMs can encode arbitrary data into a compact real-valued vector (i.e., encryption) and then losslessly decode the vector to the original message (i.e., decryption) via random subspace optimization and greedy decoding. While SELM is not amenable to conventional cryptanalysis, we investigate its security through a novel empirical variant of the classic IND-CPA (indistinguishability under chosen-plaintext attack) game and show promising results on security. Our code and datasets are available at https://github.com/OSU-NLP-Group/SELM. △ Less

Submitted 13 October, 2023; v1 submitted 15 May, 2023; originally announced May 2023.

Comments: Main text: 9 pages, 4 figures, 1 table. Work-in-progress. Project website at https://samuelstevens.me/research/encryption/

Removing Structured Noise with Diffusion Models

Authors: Tristan S. W. Stevens, Hans van Gorp, Faik C. Meral, Junseob Shin, Jason Yu, Jean-Luc Robert, Ruud J. G. van Sloun

Abstract: Solving ill-posed inverse problems requires careful formulation of prior beliefs over the signals of interest and an accurate description of their manifestation into noisy measurements. Handcrafted signal priors based on e.g. sparsity are increasingly replaced by data-driven deep generative models, and several groups have recently shown that state-of-the-art score-based diffusion models yield part… ▽ More Solving ill-posed inverse problems requires careful formulation of prior beliefs over the signals of interest and an accurate description of their manifestation into noisy measurements. Handcrafted signal priors based on e.g. sparsity are increasingly replaced by data-driven deep generative models, and several groups have recently shown that state-of-the-art score-based diffusion models yield particularly strong performance and flexibility. In this paper, we show that the powerful paradigm of posterior sampling with diffusion models can be extended to include rich, structured, noise models. To that end, we propose a joint conditional reverse diffusion process with learned scores for the noise and signal-generating distribution. We demonstrate strong performance gains across various inverse problems with structured noise, outperforming competitive baselines that use normalizing flows and adversarial networks. This opens up new opportunities and relevant practical applications of diffusion modeling for inverse problems in the context of non-Gaussian measurement models. △ Less

Submitted 17 October, 2023; v1 submitted 20 January, 2023; originally announced February 2023.

Comments: 11 pages, 7 figures, preprint

arXivEdits: Understanding the Human Revision Process in Scientific Writing

Authors: Chao Jiang, Wei Xu, Samuel Stevens

Abstract: Scientific publications are the primary means to communicate research discoveries, where the writing quality is of crucial importance. However, prior work studying the human editing process in this domain mainly focused on the abstract or introduction sections, resulting in an incomplete picture. In this work, we provide a complete computational framework for studying text revision in scientific w… ▽ More Scientific publications are the primary means to communicate research discoveries, where the writing quality is of crucial importance. However, prior work studying the human editing process in this domain mainly focused on the abstract or introduction sections, resulting in an incomplete picture. In this work, we provide a complete computational framework for studying text revision in scientific writing. We first introduce arXivEdits, a new annotated corpus of 751 full papers from arXiv with gold sentence alignment across their multiple versions of revision, as well as fine-grained span-level edits and their underlying intentions for 1,000 sentence pairs. It supports our data-driven analysis to unveil the common strategies practiced by researchers for revising their papers. To scale up the analysis, we also develop automatic methods to extract revision at document-, sentence-, and word-levels. A neural CRF sentence alignment model trained on our corpus achieves 93.8 F1, enabling the reliable matching of sentences between different versions. We formulate the edit extraction task as a span alignment problem, and our proposed method extracts more fine-grained and explainable edits, compared to the commonly used diff algorithm. An intention classifier trained on our dataset achieves 78.9 F1 on the fine-grained intent classification task. Our data and system are released at tiny.one/arxivedits. △ Less

Submitted 31 October, 2022; v1 submitted 26 October, 2022; originally announced October 2022.

Comments: This paper has been accepted to EMNLP 2022

Bootstrap** a User-Centered Task-Oriented Dialogue System

Authors: Shijie Chen, Ziru Chen, Xiang Deng, Ashley Lewis, Lingbo Mo, Samuel Stevens, Zhen Wang, Xiang Yue, Tianshu Zhang, Yu Su, Huan Sun

Abstract: We present TacoBot, a task-oriented dialogue system built for the inaugural Alexa Prize TaskBot Challenge, which assists users in completing multi-step cooking and home improvement tasks. TacoBot is designed with a user-centered principle and aspires to deliver a collaborative and accessible dialogue experience. Towards that end, it is equipped with accurate language understanding, flexible dialog… ▽ More We present TacoBot, a task-oriented dialogue system built for the inaugural Alexa Prize TaskBot Challenge, which assists users in completing multi-step cooking and home improvement tasks. TacoBot is designed with a user-centered principle and aspires to deliver a collaborative and accessible dialogue experience. Towards that end, it is equipped with accurate language understanding, flexible dialogue management, and engaging response generation. Furthermore, TacoBot is backed by a strong search engine and an automated end-to-end test suite. In bootstrap** the development of TacoBot, we explore a series of data augmentation strategies to train advanced neural language processing models and continuously improve the dialogue experience with collected real conversations. At the end of the semifinals, TacoBot achieved an average rating of 3.55/5.0. △ Less

Submitted 21 July, 2022; v1 submitted 11 July, 2022; originally announced July 2022.

Comments: Published in 1st Proceedings of Alexa Prize TaskBot (Alexa Prize 2021). TacoBot won 3rd place in the challenge. See project website https://sunlab-osu.github.io/tacobot/ for details

Accelerated Intravascular Ultrasound Imaging using Deep Reinforcement Learning

Authors: Tristan S. W. Stevens, Nishith Chennakeshava, Frederik J. de Bruijn, Martin Pekař, Ruud J. G. van Sloun

Abstract: Intravascular ultrasound (IVUS) offers a unique perspective in the treatment of vascular diseases by creating a sequence of ultrasound-slices acquired from within the vessel. However, unlike conventional hand-held ultrasound, the thin catheter only provides room for a small number of physical channels for signal transfer from a transducer-array at the tip. For continued improvement of image qualit… ▽ More Intravascular ultrasound (IVUS) offers a unique perspective in the treatment of vascular diseases by creating a sequence of ultrasound-slices acquired from within the vessel. However, unlike conventional hand-held ultrasound, the thin catheter only provides room for a small number of physical channels for signal transfer from a transducer-array at the tip. For continued improvement of image quality and frame rate, we present the use of deep reinforcement learning to deal with the current physical information bottleneck. Valuable inspiration has come from the field of magnetic resonance imaging (MRI), where learned acquisition schemes have brought significant acceleration in image acquisition at competing image quality. To efficiently accelerate IVUS imaging, we propose a framework that utilizes deep reinforcement learning for an optimal adaptive acquisition policy on a per-frame basis enabled by actor-critic methods and Gumbel top-$K$ sampling. △ Less

Submitted 24 January, 2022; originally announced January 2022.

Comments: 5 pages, 3 figures, conference

Journal ref: ICASSP 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Automated Gain Control Through Deep Reinforcement Learning for Downstream Radar Object Detection

Authors: Tristan S. W. Stevens, R. Firat Tigrek, Eric S. Tammam, Ruud J. G. van Sloun

Abstract: Cognitive radars are systems that rely on learning through interactions of the radar with the surrounding environment. To realize this, radar transmit parameters can be adapted such that they facilitate some downstream task. This paper proposes the use of deep reinforcement learning (RL) to learn policies for gain control under the object detection task. The YOLOv3 single-shot object detector is u… ▽ More Cognitive radars are systems that rely on learning through interactions of the radar with the surrounding environment. To realize this, radar transmit parameters can be adapted such that they facilitate some downstream task. This paper proposes the use of deep reinforcement learning (RL) to learn policies for gain control under the object detection task. The YOLOv3 single-shot object detector is used for the downstream task and will be concurrently used alongside the RL agent. Furthermore, a synthetic dataset is introduced which models the radar environment with use of the Grand Theft Auto V game engine. This approach allows for simulation of vast amounts of data with flexible assignment of the radar parameters to aid in the active learning process. △ Less

Submitted 8 July, 2021; originally announced July 2021.

Comments: 5 pages, 5 figures, conference

Journal ref: 2021 29th European Signal Processing Conference (EUSIPCO)

Elastic Shape Analysis of Brain Structures for Predictive Modeling of PTSD

Authors: Yuexuan Wu, Suprateek Kundu, Jennifer S. Stevens, Negar Fani, Anuj Srivastava

Abstract: There is increasing evidence on the importance of brain morphology in predicting and classifying mental disorders. However, the vast majority of current shape approaches rely heavily on vertex-wise analysis that may not successfully capture complexities of subcortical structures. Additionally, the past works do not include interactions between these structures and exposure factors. Predictive mode… ▽ More There is increasing evidence on the importance of brain morphology in predicting and classifying mental disorders. However, the vast majority of current shape approaches rely heavily on vertex-wise analysis that may not successfully capture complexities of subcortical structures. Additionally, the past works do not include interactions between these structures and exposure factors. Predictive modeling with such interactions is of paramount interest in heterogeneous mental disorders such as PTSD, where trauma exposure interacts with brain shape changes to influence behavior. We propose a comprehensive framework that overcomes these limitations by representing brain substructures as continuous parameterized surfaces and quantifying their shape differences using elastic shape metrics. Using the elastic shape metric, we compute shape summaries of subcortical data and represent individual shapes by their principal scores. These representations allow visualization tools that help localize changes when these PCs are varied. Subsequently, these PCs, the auxiliary exposure variables, and their interactions are used for regression modeling. We apply our method to data from the Grady Trauma Project, where the goal is to predict clinical measures of PTSD using shapes of brain substructures. Our analysis revealed considerably greater predictive power under the elastic shape analysis than widely used approaches such as vertex-wise shape analysis and even volumetric analysis. It helped identify local deformations in brain shapes related to change in PTSD severity. To our knowledge, this is one of the first brain shape analysis approaches that can seamlessly integrate the pre-processing steps under one umbrella for improved accuracy and are naturally able to account for interactions between brain shape and additional covariates to yield superior predictive performance when modeling clinical outcomes. △ Less

Submitted 24 May, 2021; originally announced May 2021.

Comments: 33 pages; Supplementary Materials and interactive visualizations are available in https://www.dropbox.com/home/Paper%20Interactive%20Figures

Attaining the exponent $5/4$ for the sum-product problem in finite fields

Authors: Ali Mohammadi, Sophie Stevens

Abstract: We improve the exponent in the finite field sum-product problem from $11/9$ to $5/4$, improving the results of Rudnev, Shakan and Shkredov. That is, we show that if $A\subset \mathbb{F}_p$ has cardinality $|A|\ll p^{1/2}$ then \[ \max\{|A\pm A|,|AA|\} \gtrsim |A|^\frac54 \] and \[ \max\{|A\pm A|,|A/A|\}\gtrsim |A|^\frac54\,. \] We improve the exponent in the finite field sum-product problem from $11/9$ to $5/4$, improving the results of Rudnev, Shakan and Shkredov. That is, we show that if $A\subset \mathbb{F}_p$ has cardinality $|A|\ll p^{1/2}$ then \[ \max\{|A\pm A|,|AA|\} \gtrsim |A|^\frac54 \] and \[ \max\{|A\pm A|,|A/A|\}\gtrsim |A|^\frac54\,. \] △ Less

Submitted 2 April, 2021; v1 submitted 15 March, 2021; originally announced March 2021.

Comments: 11 pages, v3 improves the exponent from 21/17 to 5/4

On sum sets of convex functions

Authors: Sophie Stevens, Audie Warren

Abstract: In this paper we prove new bounds for sums of convex or concave functions. Specifically, we prove that for all $A,B \subseteq \mathbb R$ finite sets, and for all $f,g$ convex or concave functions, we have $$|A + B|^{38}|f(A) + g(B)|^{38} \gtrsim |A|^{49}|B|^{49}.$$ This result can be used to obtain bounds on a number of two-variable expanders of interest, as well as to the asymmetric sum-produ… ▽ More In this paper we prove new bounds for sums of convex or concave functions. Specifically, we prove that for all $A,B \subseteq \mathbb R$ finite sets, and for all $f,g$ convex or concave functions, we have $$|A + B|^{38}|f(A) + g(B)|^{38} \gtrsim |A|^{49}|B|^{49}.$$ This result can be used to obtain bounds on a number of two-variable expanders of interest, as well as to the asymmetric sum-product problem. We also adjust our technique to also prove the three-variable expansion result \[ |AB+A|\gtrsim |A|^{\frac32 +\frac3{170}}\,. \] Our methods follow a series of recent developments in the sum-product literature, presenting a unified picture. Of particular interest is an adaptation of a regularisation technique of Xue, that enables us to find positive proportion subsets with certain desirable properties. △ Less

Submitted 10 February, 2021; originally announced February 2021.

Comments: 16 pages plus appendix

Low-energy decomposition results over finite fields

Authors: Ali Mohammadi, Sophie Stevens

Abstract: We prove various low-energy decomposition results, showing that we can decompose a finite set $A\subset \mathbb{F}_p$ satisfying $|A|<p^{5/8}$, into $A = S\sqcup T$ so that, for a non-degenerate quadratic $f\in \mathbb{F}_p[x,y]$, we have \[ |\{(s_1,s_2,s_3,s_4)\in S^4 : s_1 + s_2 = s_3 + s_4\}| \ll |A|^{3 - \frac15 + \varepsilon} \] and \[ |\{(t_1,t_2,t_3,t_4)\in T^4 : f(t_1, t_2) = f(t_3… ▽ More We prove various low-energy decomposition results, showing that we can decompose a finite set $A\subset \mathbb{F}_p$ satisfying $|A|<p^{5/8}$, into $A = S\sqcup T$ so that, for a non-degenerate quadratic $f\in \mathbb{F}_p[x,y]$, we have \[ |\{(s_1,s_2,s_3,s_4)\in S^4 : s_1 + s_2 = s_3 + s_4\}| \ll |A|^{3 - \frac15 + \varepsilon} \] and \[ |\{(t_1,t_2,t_3,t_4)\in T^4 : f(t_1, t_2) = f(t_3, t_4)\}|\ll |A|^{3 - \frac15 + \varepsilon}\,. \] Variations include extending this result to large $A$ and a low-energy decomposition involving additive energy of images of rational functions. This gives a quantitative improvement to a result of Roche-Newton, Shparlinski and Winterhof as well as a generalisation of a result of Rudnev, Shkredov and Stevens. We consider applications to conditional expanders, exponential sum estimates and the finite field Littlewood problem. In particular, we improve results of Mirzaei, Swaenepoel and Winterhof and Garcia. △ Less

Submitted 6 July, 2021; v1 submitted 2 February, 2021; originally announced February 2021.

Comments: 20 pages, calculation error corrected

MSC Class: 11B75

An Investigation of Language Model Interpretability via Sentence Editing

Authors: Samuel Stevens, Yu Su

Abstract: Pre-trained language models (PLMs) like BERT are being used for almost all language-related tasks, but interpreting their behavior still remains a significant challenge and many important questions remain largely unanswered. In this work, we re-purpose a sentence editing dataset, where faithful high-quality human rationales can be automatically extracted and compared with extracted model rationale… ▽ More Pre-trained language models (PLMs) like BERT are being used for almost all language-related tasks, but interpreting their behavior still remains a significant challenge and many important questions remain largely unanswered. In this work, we re-purpose a sentence editing dataset, where faithful high-quality human rationales can be automatically extracted and compared with extracted model rationales, as a new testbed for interpretability. This enables us to conduct a systematic investigation on an array of questions regarding PLMs' interpretability, including the role of pre-training procedure, comparison of rationale extraction methods, and different layers in the PLM. The investigation generates new insights, for example, contrary to the common understanding, we find that attention weights correlate well with human rationales and work better than gradient-based saliency in extracting model rationales. Both the dataset and code are available at https://github.com/samuelstevens/sentence-editing-interpretability to facilitate future interpretability research. △ Less

Submitted 26 September, 2021; v1 submitted 27 November, 2020; originally announced November 2020.

Comments: 12 pages, 9 figures. Accepted at EMNLP BlackboxNLP 2021

The Elekes-Szabó Problem and the Uniformity Conjecture

Authors: Mehdi Makhul, Oliver Roche-Newton, Sophie Stevens, Audie Warren

Abstract: In this paper we give a conditional improvement to the Elekes-Szabó problem over the rationals, assuming the Uniformity Conjecture. Our main result states that for $F\in \mathbb{Q}[x,y,z]$ belonging to a particular family of polynomials, and any finite sets $A, B, C \subset \mathbb Q$ with $|A|=|B|=|C|=n$, we have \[ |Z(F) \cap (A\times B \times C)| \ll n^{2-\frac{1}{s}}. \] The value of t… ▽ More In this paper we give a conditional improvement to the Elekes-Szabó problem over the rationals, assuming the Uniformity Conjecture. Our main result states that for $F\in \mathbb{Q}[x,y,z]$ belonging to a particular family of polynomials, and any finite sets $A, B, C \subset \mathbb Q$ with $|A|=|B|=|C|=n$, we have \[ |Z(F) \cap (A\times B \times C)| \ll n^{2-\frac{1}{s}}. \] The value of the integer $s$ is dependent on the polynomial $F$, but is always bounded by $s \leq 5$, and so even in the worst applicable case this gives a quantitative improvement on a bound of Raz, Sharir and de Zeeuw (arXiv:1504.05012). We give several applications to problems in discrete geometry and arithmetic combinatorics. For instance, for any set $P \subset \mathbb Q^2$ and any two points $p_1,p_2 \in \mathbb Q^2$, we prove that at least one of the $p_i$ satisfies the bound \[ | \{ \| p_i - p \| : p \in P \}| \gg |P|^{3/5}, \] where $\| \cdot \|$ denotes Euclidean distance. This gives a conditional improvement to a result of Sharir and Solymosi (arXiv:1308.0814). △ Less

Submitted 19 October, 2020; v1 submitted 28 September, 2020; originally announced September 2020.

Comments: A reference error has been corrected

An update on the sum-product problem

Authors: Misha Rudnev, Sophie Stevens

Abstract: We improve the best known sum-product estimates over the reals. We prove that \[ \max(|A+A|,|AA|)\geq |A|^{\frac{4}{3} + \frac{2}{1167} - o(1)}\,, \] for a finite $A\subset \mathbb R$, following a streamlining of the arguments of Solymosi, Konyagin and Shkredov. We include several new observations to our techniques. Furthermore, \[ |AA+AA|\geq |A|^{\frac{127}{80} - o(1)}\,. \] Besides, for a con… ▽ More We improve the best known sum-product estimates over the reals. We prove that \[ \max(|A+A|,|AA|)\geq |A|^{\frac{4}{3} + \frac{2}{1167} - o(1)}\,, \] for a finite $A\subset \mathbb R$, following a streamlining of the arguments of Solymosi, Konyagin and Shkredov. We include several new observations to our techniques. Furthermore, \[ |AA+AA|\geq |A|^{\frac{127}{80} - o(1)}\,. \] Besides, for a convex set $A$ we show that \[ |A+A|\geq |A|^{\frac{30}{19}-o(1)}\,. \] This paper is largely self-contained. △ Less

Submitted 2 September, 2021; v1 submitted 22 May, 2020; originally announced May 2020.

Comments: 19 pages, refereed version

Journal ref: Math. Proc. Cambridge Phil. Soc. 2021

On the Pinned Distances Problem in Positive Characteristic

Authors: Brendan Murphy, Giorgis Petridis, Thang Pham, Misha Rudnev, Sophie Stevens

Abstract: We study the Erd\H os-Falconer distance problem for a set $A\subset \mathbb{F}^2$, where $\mathbb{F}$ is a field of positive characteristic $p$. If $\mathbb{F}=\mathbb{F}_p$ and the cardinality $|A|$ exceeds $p^{5/4}$, we prove that $A$ determines an asymptotically full proportion of the feasible $p$ distances. For small sets $A$, namely when $|A|\leq p^{4/3}$ over any $\mathbb{F}$, we prove that… ▽ More We study the Erd\H os-Falconer distance problem for a set $A\subset \mathbb{F}^2$, where $\mathbb{F}$ is a field of positive characteristic $p$. If $\mathbb{F}=\mathbb{F}_p$ and the cardinality $|A|$ exceeds $p^{5/4}$, we prove that $A$ determines an asymptotically full proportion of the feasible $p$ distances. For small sets $A$, namely when $|A|\leq p^{4/3}$ over any $\mathbb{F}$, we prove that either $A$ determines $\gg|A|^{2/3}$. For both large and small sets, the results proved are in fact for pinned distances. △ Less

Submitted 8 July, 2021; v1 submitted 1 March, 2020; originally announced March 2020.

Comments: A substantial revision of the older version. The main results are unchanged. arXiv admin note: text overlap with arXiv:1908.04618

MSC Class: 52C10; 11E88; 11T23; 14N10; 52C35; 51E20

Journal ref: J. Lond. Math. Soc.(2) 105 (1), 469-499, 2022

Bisector energy and pinned distances in positive characteristic

Authors: Brendan Murphy, Misha Rudnev, Sophie Stevens

Abstract: We prove a new lower bound for the number of pinned distances over finite fields: if $A$ is a sufficiently small subset of $\mathbb{F}_q^2$, then there is an element in $A$ that determines $\gg |A|^{2/3}$ distinct distances to other elements of $A$. Combined with results for large subsets $A\subseteq\mathbb{F}_q^2$, this improves all previously known lower bounds on distinct distances over finite… ▽ More We prove a new lower bound for the number of pinned distances over finite fields: if $A$ is a sufficiently small subset of $\mathbb{F}_q^2$, then there is an element in $A$ that determines $\gg |A|^{2/3}$ distinct distances to other elements of $A$. Combined with results for large subsets $A\subseteq\mathbb{F}_q^2$, this improves all previously known lower bounds on distinct distances over finite fields. In fact, we obtain an upper bound for the number of isosceles triangles determined by $A$. For that we use the concept of bisector energy. It turns out that the latter can be expressed as a point-plane incidence bound, so one can use a theorem of the third author. The conversion to this incidence problem relies on the Blaschke-Grünwald kinematic map** -- an embedding of the group of rigid motions of $\mathbb{F}_q^2$ into an open subset of the projective three space. This has long been known in kinematics and geometric algebra; we provide a proof for arbitrary fields using Clifford algebras. △ Less

Submitted 1 November, 2019; v1 submitted 13 August, 2019; originally announced August 2019.

Comments: 17 pages. Updated in response to comments by Giorgis Petridis and Thang Pham. Removed alternate proofs for existing results for large sets

MSC Class: 52C10; 11E88; 11T23; 14N10; 52C35; 51E20

The isospin and neutron-to-proton excess dependence of short-range correlations

Authors: Jan Ryckebusch, Wim Cosyn, Sam Stevens, Corneel Casert, Jannes Nys

Abstract: We provide a systematic study of the isospin composition and neutron-to-proton $\left( \frac{N}{Z} \right)$ ratio dependence of nuclear short-range correlations (SRC) across the nuclear mass table. We use the low-order correlation operator approximation (LCA) to compute the SRC contribution to the single-nucleon momentum distributions for 14 different nuclei from $A=4$ to $A=208$. Ten asymmetric n… ▽ More We provide a systematic study of the isospin composition and neutron-to-proton $\left( \frac{N}{Z} \right)$ ratio dependence of nuclear short-range correlations (SRC) across the nuclear mass table. We use the low-order correlation operator approximation (LCA) to compute the SRC contribution to the single-nucleon momentum distributions for 14 different nuclei from $A=4$ to $A=208$. Ten asymmetric nuclei are included for which the neutrons outnumber the protons by a factor of up to 1.54. The computed momentum distributions are used to extract the pair composition of the SRC. We find that there is a comprehensive picture for the isospin composition of SRC and their evolution with nucleon momentum. We also compute the non-relativistic kinetic energy of neutrons and protons and its evolution with nuclear mass $A$ and $\frac{N}{Z}$. Confirming the conclusions from alternate studies it is shown that the minority species (protons) become increasingly more short-range correlated as the neutron-to-proton ratio increases. We forge connections between measured nucleon-knockout quantities sensitive to SRC and single-nucleon momentum distributions. It is shown that the LCA can account for the observed trends in the data, like the fact that in neutron-rich nuclei the protons are responsible for an unexpectedly large fraction of the high-momentum components. △ Less

Submitted 15 March, 2019; v1 submitted 29 August, 2018; originally announced August 2018.

Comments: 8 pages, 7 figures

Galois self-dual cuspidal types and Asai local factors

Authors: U. K. Anandavardhanan, Robert Kurinczuk, Nadir Matringe, Vincent Sécherre, Shaun Stevens

Abstract: Let $F/F_{\mathsf{o}}$ be a quadratic extension of non-archimedean locally compact fields of odd residual characteristic and $σ$ be its non-trivial automorphism. We show that any $σ$-self-dual cuspidal representation of ${\rm GL}_n(F)$ contains a $σ$-self-dual Bushnell--Kutzko type. Using such a type, we construct an explicit test vector for Flicker's local Asai $L$-function of a… ▽ More Let $F/F_{\mathsf{o}}$ be a quadratic extension of non-archimedean locally compact fields of odd residual characteristic and $σ$ be its non-trivial automorphism. We show that any $σ$-self-dual cuspidal representation of ${\rm GL}_n(F)$ contains a $σ$-self-dual Bushnell--Kutzko type. Using such a type, we construct an explicit test vector for Flicker's local Asai $L$-function of a ${\rm GL}_n(F_{\mathsf{o}})$-distinguished cuspidal representation and compute the associated Asai root number. Finally, by using global methods, we compare this root number to Langlands--Shahidi's local Asai root number, and more generally we compare the corresponding epsilon factors for any cuspidal representation. △ Less

Submitted 18 April, 2019; v1 submitted 20 July, 2018; originally announced July 2018.

Comments: 61 pages, final version to appear in JEMS

MSC Class: 22E50; 11F70

Probing short-range correlations in asymmetric nuclei with quasi-free pair knockout reactions

Authors: Sam Stevens, Jan Ryckebusch, Wim Cosyn, Andreas Waets

Abstract: Short-range correlations (SRC) in asymmetric nuclei with an unusual neutron-to-proton ratio can be studied with quasi-free two-nucleon knockout processes following the collision between accelerated ions and a proton target. We derive an approximate factorized cross section for those SRC-driven $p(A,p^{\prime} N_1 N_2)$ reactions. Our reaction model hinges on the factorization properties of SRC-dri… ▽ More Short-range correlations (SRC) in asymmetric nuclei with an unusual neutron-to-proton ratio can be studied with quasi-free two-nucleon knockout processes following the collision between accelerated ions and a proton target. We derive an approximate factorized cross section for those SRC-driven $p(A,p^{\prime} N_1 N_2)$ reactions. Our reaction model hinges on the factorization properties of SRC-driven $A(e, e^\prime N_1 N_2)$ reactions for which strong indications are found in theory-experiment comparisons. In order to put our model to the test we compare its predictions with results of $^{12}\text{C}(p,p^{\prime} pn)$ measurements conducted at Brookhaven National Laboratory (BNL) and find a fair agreement. The model can also reproduce characteristic features of SRC-driven two-nucleon knockout reactions, like back-to-back emission of the correlated nucleons. We study the asymmetry dependence of nuclear SRC by providing predictions for the ratio of proton-proton to proton-neutron knockout cross sections for the carbon isotopes $^{9-15}$C thereby covering neutron excess values $(N-Z)/Z$ between -0.5 and +0.5. △ Less

Submitted 8 January, 2018; v1 submitted 18 July, 2017; originally announced July 2017.

Comments: 20 pages, 5 figures. Final version published in Physics Letters B, minor changes in the text

Journal ref: Phys.Lett. B777 (2018) 374-380

Jordan blocks of cuspidal representations of symplectic groups

Authors: Corinne Blondel, Guy Henniart, Shaun Stevens

Abstract: Let $G$ be a symplectic group over a nonarchimedean local field of characteristic zero and odd residual characteristic. Given an irreducible cuspidal representation of G, we determine its Langlands parameter (equivalently, its Jordan blocks in the language of Moeglin) in terms of the local data from which the representation is explicitly constructed, up to a possible unramified twist in each block… ▽ More Let $G$ be a symplectic group over a nonarchimedean local field of characteristic zero and odd residual characteristic. Given an irreducible cuspidal representation of G, we determine its Langlands parameter (equivalently, its Jordan blocks in the language of Moeglin) in terms of the local data from which the representation is explicitly constructed, up to a possible unramified twist in each block of the parameter. We deduce a Ramification Theorem for $G$, giving a bijection between the set of endo-parameters for $G$ and the set of restrictions to wild inertia of discrete Langlands parameters for $G$, compatible with the local Langlands correspondence. The main tool consists in analysing the intertwining Hecke algebra of a good cover, in the sense of Bushnell--Kutzko, for parabolic induction from a cuspidal representation of $G\times\mathrm{GL}_n$, seen as a maximal Levi subgroup of a bigger symplectic group, in order to determine its (ir)reducibility; a criterion of Moeglin then relates this to Langlands parameters. △ Less

Submitted 11 April, 2017; originally announced April 2017.

Comments: 64 pages

MSC Class: 22E50; 11F70

Journal ref: Alg. Number Th. 12 (2018) 2327-2386

On depth zero L-packets for classical groups

Authors: Jaime Lust, Shaun Stevens

Abstract: By computing reducibility points of parabolically induced representations, we construct, to within at most two unramified quadratic characters, the Langlands parameter of an arbitrary depth zero irreducible cuspidal representation $π$ of a classical group (which may be not-quasi-split) over a nonarchimedean local field of odd residual characteristic. From this, we can explicitly describe all the i… ▽ More By computing reducibility points of parabolically induced representations, we construct, to within at most two unramified quadratic characters, the Langlands parameter of an arbitrary depth zero irreducible cuspidal representation $π$ of a classical group (which may be not-quasi-split) over a nonarchimedean local field of odd residual characteristic. From this, we can explicitly describe all the irreducible cuspidal representations in the union of one, two, or four L-packets, containing $π$. These results generalize the work of DeBacker-Reeder (in the case of classical groups) from regular to arbitrary tame Langlands parameters. △ Less

Submitted 25 November, 2016; originally announced November 2016.

Comments: 36 pages

MSC Class: 22F50

The regular representations of $\mathrm{GL}_{N}$ over finite local principal ideal rings

Authors: Alexander Stasinski, Shaun Stevens

Abstract: Let $\mathfrak{o}$ be the ring of integers in a non-Archimedean local field with finite residue field, $\mathfrak{p}$ its maximal ideal, and $r\geq2$ an integer. An irreducible representation of the finite group $G_{r}=\mathrm{GL}_{N}(\mathfrak{o}/\mathfrak{p}^{r})$ is called regular if its restriction to the principal congruence kernel… ▽ More Let $\mathfrak{o}$ be the ring of integers in a non-Archimedean local field with finite residue field, $\mathfrak{p}$ its maximal ideal, and $r\geq2$ an integer. An irreducible representation of the finite group $G_{r}=\mathrm{GL}_{N}(\mathfrak{o}/\mathfrak{p}^{r})$ is called regular if its restriction to the principal congruence kernel $K^{r-1}=1+\mathfrak{p}^{r-1}\mathrm{M}_{N}(\mathfrak{o}/\mathfrak{p}^{r})$ consists of representations whose stabilisers modulo $K^{1}$ are centralisers of regular elements in $\mathrm{M}_{N}(\mathfrak{o}/\mathfrak{p})$. The regular representations form the largest class of representations of $G_{r}$ which is currently amenable to explicit construction. Their study, motivated by constructions of supercuspidal representations, goes back to Shintani, but the general case remained open for a long time. In this paper we give an explicit construction of all the regular representations of $G_{r}$. △ Less

Submitted 15 November, 2016; originally announced November 2016.

Comments: 20 pages

Towards an explicit local Jacquet-Langlands correspondence beyond the cuspidal case

Authors: Vincent Sécherre, Shaun Stevens

Abstract: We show how the modular representation theory of inner forms of general linear groups over a non-Archimedean local field can be brought to bear on the complex theory in a remarkable way. Let F be a non-Archimedean locally compact field of residue characteristic p, and let G be an inner form of the general linear group GL(n,F). We consider the problem of describing explicitly the local Jacquet--Lan… ▽ More We show how the modular representation theory of inner forms of general linear groups over a non-Archimedean local field can be brought to bear on the complex theory in a remarkable way. Let F be a non-Archimedean locally compact field of residue characteristic p, and let G be an inner form of the general linear group GL(n,F). We consider the problem of describing explicitly the local Jacquet--Langlands correspondence between the complex discrete series representations of G and GL(n,F), in terms of type theory. We show that the congruence properties of the local Jacquet--Langlands correspondence exhibited by A. Mínguez and the first named author give information about the explicit description of this correspondence. We prove that the problem of the invariance of the endo-class by the Jacquet--Langlands correspondence can be reduced to the case where the representations $π$ and its Jacquet--Langlands transfer JL($π$) are both cuspidal with torsion number 1. We also give an explicit description of the Jacquet--Langlands correspondence for all essentially tame discrete series representations of G, up to an unramified twist, in terms of admissible pairs, generalizing previous results by Bushnell and Henniart. In positive depth, our results are the first beyond the case where $π$ and JL($π$) are both cuspidal. △ Less

Submitted 14 November, 2016; originally announced November 2016.

Journal ref: Compositio Math. 155 (2019) 1853-1887

Endo-parameters for p-adic classical groups

Authors: Robert Kurinczuk, Daniel Skodlerack, Shaun Stevens

Abstract: For a classical group over a non-archimedean local field of odd residual characteristic p, we prove that two cuspidal types, defined over an algebraically closed field C of characteristic different from p, intertwine if and only if they are conjugate. This completes work of the first and third authors who showed that every irreducible cuspidal C-representation of a classical group is compactly ind… ▽ More For a classical group over a non-archimedean local field of odd residual characteristic p, we prove that two cuspidal types, defined over an algebraically closed field C of characteristic different from p, intertwine if and only if they are conjugate. This completes work of the first and third authors who showed that every irreducible cuspidal C-representation of a classical group is compactly induced from a cuspidal type. We generalize Bushnell and Henniart's notion of endo-equivalence to semisimple characters of general linear groups and to self-dual semisimple characters of classical groups, and introduce (self-dual) endo-parameters. We prove that these parametrize intertwining classes of (self-dual) semisimple characters and conjecture that they are in bijection with wild Langlands parameters, compatibly with the local Langlands correspondence. △ Less

Submitted 31 August, 2020; v1 submitted 8 November, 2016; originally announced November 2016.

Comments: 81 pages, to appear in Inventiones Mathematicae

MSC Class: 22E50; 11F70

An Improved Point-Line Incidence Bound Over Arbitrary Fields

Authors: Sophie Stevens, Frank de Zeeuw

Abstract: We prove a new upper bound for the number of incidences between points and lines in a plane over an arbitrary field $\mathbb{F}$, a problem first considered by Bourgain, Katz and Tao. Specifically, we show that $m$ points and $n$ lines in $\mathbb{F}^2$, with $m^{7/8}<n<m^{8/7}$, determine at most $O(m^{11/15}n^{11/15})$ incidences (where, if $\mathbb{F}$ has positive characteristic $p$, we assume… ▽ More We prove a new upper bound for the number of incidences between points and lines in a plane over an arbitrary field $\mathbb{F}$, a problem first considered by Bourgain, Katz and Tao. Specifically, we show that $m$ points and $n$ lines in $\mathbb{F}^2$, with $m^{7/8}<n<m^{8/7}$, determine at most $O(m^{11/15}n^{11/15})$ incidences (where, if $\mathbb{F}$ has positive characteristic $p$, we assume $m^{-2}n^{13}\ll p^{15}$). This improves on the previous best known bound, due to Jones. To obtain our bound, we first prove an optimal point-line incidence bound on Cartesian products, using a reduction to a point-plane incidence bound of Rudnev. We then cover most of the point set with Cartesian products, and we bound the incidences on each product separately, using the bound just mentioned. We give several applications, to sum-product-type problems, an expander problem of Bourgain, the distinct distance problem and Beck's theorem. △ Less

Submitted 18 July, 2017; v1 submitted 20 September, 2016; originally announced September 2016.

Comments: 18 pages. To appear in the Bulletin of the London Mathematical Society

On The Energy Variant of the Sum-Product Conjecture

Authors: Misha Rudnev, Ilya D. Shkredov, Sophie Stevens

Abstract: We prove new exponents for the energy version of the Erdős-Szemerédi sum-product conjecture, raised by Balog and Wooley. They match the previously established milestone values for the standard formulation of the question, both for general fields and the special case of real or complex numbers, and appear to be the best ones attainable within the currently available technology. Further results are… ▽ More We prove new exponents for the energy version of the Erdős-Szemerédi sum-product conjecture, raised by Balog and Wooley. They match the previously established milestone values for the standard formulation of the question, both for general fields and the special case of real or complex numbers, and appear to be the best ones attainable within the currently available technology. Further results are obtained about multiplicative energies of additive shifts and a strengthened energy version of the "few sums, many products" inequality of Elekes and Ruzsa. The latter inequality enables us to obtain a minor improvement of the state-of the art sum-product exponent over the reals due to Konyagin and the second author, up to $\frac{4}{3}+\frac{1}{1509}$. An application of energy estimates to an instance of arithmetic growth in prime residue fields is presented. △ Less

Submitted 5 June, 2017; v1 submitted 18 July, 2016; originally announced July 2016.

Comments: 25 pages. A new best sum-product exponent is attained over the reals via a new corollary which locally improves the two recent sum-product papers by Konyagin and Shkredov

MSC Class: 68R05; 11B75

Cuspidal $\ell$-modular representations of $p$-adic classical groups

Authors: Robert Kurinczuk, Shaun Stevens

Abstract: For a classical group over a non-archimedean local field of odd residual characteristic p, we construct all cuspidal representations over an arbitrary algebraically closed field of characteristic different from p, as representations induced from a cuspidal type. We also give a fundamental step towards the classification of cuspidal representations, identifying when certain cuspidal types induce to… ▽ More For a classical group over a non-archimedean local field of odd residual characteristic p, we construct all cuspidal representations over an arbitrary algebraically closed field of characteristic different from p, as representations induced from a cuspidal type. We also give a fundamental step towards the classification of cuspidal representations, identifying when certain cuspidal types induce to equivalent representations; this result is new even in the case of complex representations. Finally, we prove that the representations induced from more general types are quasi-projective, a crucial tool for extending the results here to arbitrary irreducible representations. △ Less

Submitted 26 November, 2015; v1 submitted 7 September, 2015; originally announced September 2015.

Generalized Local Coefficients

Authors: Carlos De la Mora, Shaun Stevens

Abstract: In this paper we showed that under two assumptions we are able to define interesting functions that we call generalized local coefficients. We showed that in the quasi-split case generalized local coefficients are up to a positive constant the same as Shahidi's local coefficients. We provide a proof that the non quasi-split group $GL_m(D)$, for a central division algebra $D$ satisfies those assump… ▽ More In this paper we showed that under two assumptions we are able to define interesting functions that we call generalized local coefficients. We showed that in the quasi-split case generalized local coefficients are up to a positive constant the same as Shahidi's local coefficients. We provide a proof that the non quasi-split group $GL_m(D)$, for a central division algebra $D$ satisfies those assumptions. We also showed that generalized local coefficients satisfy nice properties, like the relation to Plancherel measures and multiplicativity inherited by that of intertwining operators. Generalized local coefficients are only defined for representations that are $(Y,\varphi)$-generic which is a generalization of generic representations in the quasi-split case. Here $Y$ denotes a nilpotent element in the Lie algebra of the group and $\varphi$ is a co-character related to $Y$. △ Less

Submitted 29 June, 2015; originally announced June 2015.

Comments: 23 pages

Towards the Jacquet Conjecture on the Local Converse Problem for $p$-adic $\mathrm{GL}_n$

Authors: Dihua Jiang, Chufeng Nien, Shaun Stevens

Abstract: The Local Converse Problem is to determine how the family of the local gamma factors $γ(s,π\timesτ,ψ)$ characterizes the isomorphism class of an irreducible admissible generic representation $π$ of $\mathrm{GL}_n(F)$, with $F$ a non-archimedean local field, where $τ$ runs through all irreducible supercuspidal representations of $\mathrm{GL}_r(F)$ and $r$ runs through positive integers. The Jacquet… ▽ More The Local Converse Problem is to determine how the family of the local gamma factors $γ(s,π\timesτ,ψ)$ characterizes the isomorphism class of an irreducible admissible generic representation $π$ of $\mathrm{GL}_n(F)$, with $F$ a non-archimedean local field, where $τ$ runs through all irreducible supercuspidal representations of $\mathrm{GL}_r(F)$ and $r$ runs through positive integers. The Jacquet conjecture asserts that it is enough to take $r=1,2,\ldots,\left[\frac{n}{2}\right]$. Based on arguments in the work of Henniart and of Chen giving preliminary steps towards the Jacquet conjecture, we formulate a general approach to prove the Jacquet conjecture. With this approach, the Jacquet conjecture is proved under an assumption which is then verified in several cases, including the case of level zero representations. △ Less

Submitted 10 April, 2015; originally announced April 2015.

Comments: Authors' final version before typesetting; first published in the Journal of the European Mathematical Society in volume 17 (2015), published by the European Mathematical Society

MSC Class: Primary 11F70; 22E50; Secondary 11F85; 22E55

Journal ref: J. Eur. Math. Soc. 17 (2015), 991-1007

Intertwining semisimple characters for p-adic classical groups

Authors: Daniel Skodlerack, Shaun Stevens

Abstract: Let~$G$ be a unitary group of an~$ε$-hermitian form~$h$ given over a nonarchimedean local field~$F_0$ of odd residue characteristic. We introduce a geometric combinatoric condition under which we prove "Intertwining implies Conjugacy" for semisimple characters of~$G$ and the general linear group of the ambient vector space of~$G$. Further we prove a Skolem-Noether result for the action of~$G$ on i… ▽ More Let~$G$ be a unitary group of an~$ε$-hermitian form~$h$ given over a nonarchimedean local field~$F_0$ of odd residue characteristic. We introduce a geometric combinatoric condition under which we prove "Intertwining implies Conjugacy" for semisimple characters of~$G$ and the general linear group of the ambient vector space of~$G$. Further we prove a Skolem-Noether result for the action of~$G$ on its Lie algebra, more precisely two Lie algebra elements of~$G$ which have the same characteristic polynomial over~$F$ must be conjugate under an element of~$G$ if there are corresponding semisimple characters which intertwine over an element of~$G$ } Let~$G$ be a unitary group over a nonarchimedean local field of odd residual characteristic. This paper concerns the study of the "wild part" of the irreducible smooth representations of~$G$, encoded in a so-called "semisimple character". We prove two fundamental results concerning them, which are crucial steps towards a classification of the cuspidal representations of~$G$. First we introduce a geometric combinatoric condition under which we prove an "intertwining implies conjugacy" theorem for semisimple characters, both in~$G$ and in the ambient general linear group. Second, we prove a Skolem--Noether theorem for the action of~$G$ on its Lie algebra; more precisely, two semisimple elements of the Lie algebra of~$G$ which have the same characteristic polynomial must be conjugate under an element of~$G$ if there are corresponding semisimple strata which are intertwined by an element of~$G$. △ Less

Submitted 8 November, 2016; v1 submitted 30 March, 2015; originally announced March 2015.

Comments: submitted to Compostio Mathematica

Patterns in Illinois Educational School Data

Authors: Cacey S. Stevens, Michael P. Marder, Sidney R. Nagel

Abstract: We examine Illinois educational data from standardized exams and analyze primary factors affecting the achievement of public school students. We focus on the simplest possible models: representation of data through visualizations and regressions on single variables. Exam scores are shown to depend on school type, location, and poverty concentration. For most schools in Illinois, student test score… ▽ More We examine Illinois educational data from standardized exams and analyze primary factors affecting the achievement of public school students. We focus on the simplest possible models: representation of data through visualizations and regressions on single variables. Exam scores are shown to depend on school type, location, and poverty concentration. For most schools in Illinois, student test scores decline linearly with poverty concentration. However Chicago must be treated separately. Selective schools in Chicago, as well as some traditional and charter schools, deviate from this pattern based on poverty. For any poverty level, Chicago schools perform better than those in the rest of Illinois. Selective programs for gifted students show high performance at each grade level, most notably at the high school level, when compared to other Illinois school types. The case of Chicago charter schools is more complex. In the last six years, their students' scores overtook those of students in traditional Chicago high schools. △ Less

Submitted 29 January, 2015; originally announced February 2015.

Comments: 9 pages, 6 figures

On the Jacquet Conjecture on the Local Converse Problem for p-adic GL_n

Authors: Moshe Adrian, Baiying Liu, Shaun Stevens, Peng Xu

Abstract: Based on previous results of Jiang, Nien and the third author, we prove that any two minimax unitarizable supercuspidals of GL_N that have the same depth and central character admit a special pair of Whittaker functions. This result gives a new reduction towards a final proof of Jacquet's conjecture on the local converse problem for GL_N. As a corollary of our result, we prove Jacquet's conjecture… ▽ More Based on previous results of Jiang, Nien and the third author, we prove that any two minimax unitarizable supercuspidals of GL_N that have the same depth and central character admit a special pair of Whittaker functions. This result gives a new reduction towards a final proof of Jacquet's conjecture on the local converse problem for GL_N. As a corollary of our result, we prove Jacquet's conjecture for GL_N, when N is prime. △ Less

Submitted 16 September, 2014; originally announced September 2014.

Comments: 15 pages

MSC Class: 22E50

Halving dynamical systems

Authors: Shaun Stevens, Tom Ward, Stefanie Zegowitz

Abstract: The relationship between two dynamical systems, one of which is obtained from the other by forming the quotient by an action of an involution commuting with the dynamics, is studied. The constraints and the possible extent of freedom in the relationship between the growth of closed orbits in pairs of systems related in this way is explored. The relationship between two dynamical systems, one of which is obtained from the other by forming the quotient by an action of an involution commuting with the dynamics, is studied. The constraints and the possible extent of freedom in the relationship between the growth of closed orbits in pairs of systems related in this way is explored. △ Less

Submitted 29 July, 2014; originally announced July 2014.

MSC Class: 37P35

Journal ref: Contemp. Mathematics, 669, 285-298 (2016)

Scaling of the Splash Threshold for Low-Viscosity Fluids

Authors: Cacey S. Stevens

Abstract: The ambient gas pressure is determined for the onset of splashing of low-viscosity liquid drops on smooth dry surfaces as we change the control parameters: drop impact velocity, drop radius, viscosity, surface tension, density, and gas molecular weight. This threshold pressure indicates that there are two distinct regimes when drop impact velocity is varied. By rescaling data using functions of on… ▽ More The ambient gas pressure is determined for the onset of splashing of low-viscosity liquid drops on smooth dry surfaces as we change the control parameters: drop impact velocity, drop radius, viscosity, surface tension, density, and gas molecular weight. This threshold pressure indicates that there are two distinct regimes when drop impact velocity is varied. By rescaling data using functions of only three dimensionless numbers, the commonly used Reynolds and Weber numbers, as well as the ratio of drop radius to gas mean free path, all data is collapsed to a single curve that encompasses both regimes. △ Less

Submitted 12 March, 2014; originally announced March 2014.

Comments: 5 pages, 5 figures

Comparison of splashing in high and low viscosity liquids

Authors: Cacey S. Stevens, Andrzej Latka, Sidney R. Nagel

Abstract: We explore the evolution of a splash when a liquid drop impacts a smooth, dry surface. There are two splashing regimes that occur when the liquid viscosity is varied, as is evidenced by its dependence on ambient gas pressure. A high-viscosity drop splashes by emitting a thin sheet of liquid from a spreading liquid lamella long after the drop has first contacted the solid. Likewise, we find that th… ▽ More We explore the evolution of a splash when a liquid drop impacts a smooth, dry surface. There are two splashing regimes that occur when the liquid viscosity is varied, as is evidenced by its dependence on ambient gas pressure. A high-viscosity drop splashes by emitting a thin sheet of liquid from a spreading liquid lamella long after the drop has first contacted the solid. Likewise, we find that there is also a delay in the ejection of a thin sheet when a low-viscosity drop splashes. We show how the ejection time of the thin sheet depends on liquid viscosity and ambient gas pressure. △ Less

Submitted 26 February, 2014; originally announced February 2014.

Comments: 4 pages, 5 figures

Block decomposition of the category of l-modular smooth representations of GL(n,F) and its inner forms

Authors: Vincent Sécherre, Shaun Stevens

Abstract: Let F be a non-Archimedean locally compact field of residue characteristic p, let D be a finite dimensional central division F-algebra and let R be an algebraically closed field of characteristic different from p. To any irreducible smooth representation of G=GL(m,D) with coefficients in R, we can attach a uniquely determined inertial class of supercuspidal pairs of G. This provides us with a part… ▽ More Let F be a non-Archimedean locally compact field of residue characteristic p, let D be a finite dimensional central division F-algebra and let R be an algebraically closed field of characteristic different from p. To any irreducible smooth representation of G=GL(m,D) with coefficients in R, we can attach a uniquely determined inertial class of supercuspidal pairs of G. This provides us with a partition of the set of all isomorphism classes of irreducible representations of G. We write R(G) for the category of all smooth representations of G with coefficients in R. To any inertial class O of supercuspidal pairs of G, we can attach the subcategory R(O) made of smooth representations all of whose irreducible subquotients are in the subset determined by this inertial class. We prove that R(G) decomposes into the product of the R(O), where O ranges over all possible inertial class of supercuspidal pairs of G, and that each summand R(O) is indecomposable. △ Less

Submitted 21 February, 2014; originally announced February 2014.

Comments: 37 pages

MSC Class: 22E50

A study of volatile compounds in the breath of children with type 1 diabetes

Authors: S Stevens, C Garner, C Wei, R Greenwood, J Hamilton-Shield, B de Lacy Costello, N Ratcliffe, C Probert

Abstract: A pilot study of exhaled volatile compounds and their correlation with blood glucose levels in eight children with type 1 diabetes is reported. Five paired blood and breath samples were obtained from each child over a 6 hour period. The blood glucose concentration ranged from 41.4 to 435.6 mg/dL. Breath samples were collected in Tedlar bags and immediately evacuated through thermal desorption tube… ▽ More A pilot study of exhaled volatile compounds and their correlation with blood glucose levels in eight children with type 1 diabetes is reported. Five paired blood and breath samples were obtained from each child over a 6 hour period. The blood glucose concentration ranged from 41.4 to 435.6 mg/dL. Breath samples were collected in Tedlar bags and immediately evacuated through thermal desorption tubes packed with Carbopack B and C. The VOCs were later recovered by thermal desorption and analysed using gas chromatography mass spectrometry. The study identified 74 volatile compounds present in at least 10% of the patient samples. Of these 74 volatiles 36 were found in all patient samples tested. Further analysis of the 36 compounds found that none showed significant overall correlation with blood glucose levels. Isoprene showed a weak negative correlation with blood glucose levels. Acetone was found to have no correlation with blood glucose levels for the patients studied. Some patients showed significant individual correlation between the relative peak areas of certain compounds and blood glucose levels. However, there was no consistent pattern observed within these results across all 8 patients. Additional breath samples were collected in Tedlar bags and analysed using SIFTMS for 3 of the patients and a healthy control. The levels of 24 volatiles are reported and were found to be generally consistent with previously reported SIFT-MS data. In agreement with the GCMS data, no compounds exhibited a significant overall correlation with blood glucose level. △ Less

Submitted 18 March, 2013; originally announced March 2013.

Comments: 18 pages

Semisimple types for p-adic classical groups

Authors: Michitaka Miyauchi, Shaun Stevens

Abstract: We construct, for any symplectic, unitary or special orthogonal group over a locally compact nonarchimedean local field of odd residual characteristic, a type for each Bernstein component of the category of smooth representations, using Bushnell--Kutzko's theory of covers. Moreover, for a component corresponding to a cuspidal representation of a maximal Levi subgroup, we prove that the Hecke algeb… ▽ More We construct, for any symplectic, unitary or special orthogonal group over a locally compact nonarchimedean local field of odd residual characteristic, a type for each Bernstein component of the category of smooth representations, using Bushnell--Kutzko's theory of covers. Moreover, for a component corresponding to a cuspidal representation of a maximal Levi subgroup, we prove that the Hecke algebra is either abelian, or a generic Hecke algebra on an infinite dihedral group, with parameters which are, at least in principle, computable via results of Lusztig. △ Less

Submitted 3 December, 2012; originally announced December 2012.

Comments: 37 pges

MSC Class: 22E50

Creation of prompt and thin-sheet splashing by varying surface roughness or increasing air pressure

Authors: Andrzej Latka, Ariana Strandburg-Peshkin, Michelle M. Driscoll, Cacey S. Stevens, Sidney R. Nagel

Abstract: A liquid drop impacting a solid surface may splash by emitting a thin liquid sheet that subsequently breaks apart or by promptly ejecting droplets from the advancing liquid-solid contact line. Using high-speed imaging, we show that air pressure and surface roughness influence both splash mechanisms. Roughness increases prompt splashing at the advancing contact line but inhibits the formation of th… ▽ More A liquid drop impacting a solid surface may splash by emitting a thin liquid sheet that subsequently breaks apart or by promptly ejecting droplets from the advancing liquid-solid contact line. Using high-speed imaging, we show that air pressure and surface roughness influence both splash mechanisms. Roughness increases prompt splashing at the advancing contact line but inhibits the formation of the thin sheet. If the air pressure is lowered, droplet ejection is suppressed not only during thin-sheet formation but for prompt splashing as well. The threshold pressure depends on impact velocity, liquid viscosity and surface roughness. △ Less

Submitted 22 June, 2012; v1 submitted 13 March, 2012; originally announced March 2012.