-
Bifurcation diagrams in a class of Kolmogorov systems
Authors:
G. Moza,
C. Lazureanu,
F. Munteanu,
C. Sterbeti,
A. Florea
Abstract:
We study a two-dimensional Kolmogorov system when its two parameters vary in a small neighbourhood of the value $0.$ The local behavior of the system is described in terms of bifurcation diagrams.
We study a two-dimensional Kolmogorov system when its two parameters vary in a small neighbourhood of the value $0.$ The local behavior of the system is described in terms of bifurcation diagrams.
△ Less
Submitted 21 March, 2024;
originally announced March 2024.
-
Analysis of a class of Kolmogorov systems
Authors:
G. Moza,
C. Lazureanu,
F. Munteanu,
C. Sterbeti,
A. Florea
Abstract:
A two-dimensional Kolmogorov system with two parameters and having a degenerate condition is studied in this work. We obtain local analytical properties of the system when the parameters vary in a sufficiently small neighborhood of the origin. The behavior of the system is described by bifurcation diagrams. Applications of Kolmogorov systems can be found particularly in modeling population dynamic…
▽ More
A two-dimensional Kolmogorov system with two parameters and having a degenerate condition is studied in this work. We obtain local analytical properties of the system when the parameters vary in a sufficiently small neighborhood of the origin. The behavior of the system is described by bifurcation diagrams. Applications of Kolmogorov systems can be found particularly in modeling population dynamics in biology and ecology.
△ Less
Submitted 20 March, 2024;
originally announced March 2024.
-
Negative discrete moments of the derivative of the Riemann zeta-function
Authors:
Hung M. Bui,
Alexandra Florea,
Micah B. Milinovich
Abstract:
We obtain conditional upper bounds for negative discrete moments of the derivative of the Riemann zeta-function averaged over a subfamily of zeros of the zeta function which is expected to have full density inside the set of all zeros. For $k\leq 1/2$, our bounds for the $2k$-th moments are expected to be almost optimal. Assuming a conjecture about the maximum size of the argument of the zeta func…
▽ More
We obtain conditional upper bounds for negative discrete moments of the derivative of the Riemann zeta-function averaged over a subfamily of zeros of the zeta function which is expected to have full density inside the set of all zeros. For $k\leq 1/2$, our bounds for the $2k$-th moments are expected to be almost optimal. Assuming a conjecture about the maximum size of the argument of the zeta function on the critical line, we obtain upper bounds for these negative moments of the same strength while summing over a larger subfamily of zeta zeros.
△ Less
Submitted 5 October, 2023;
originally announced October 2023.
-
Moments of Artin-Schreier L-functions
Authors:
Alexandra Florea,
Edna Jones,
Matilde Lalin
Abstract:
We compute moments of $L$-functions associated to the polynomial, odd polynomial and ordinary families of Artin-Schreier covers over $\mathbb{F}_q$, where $q$ is a power of a prime $p$, when the size of the finite field is fixed and the genus of the family goes to infinity. In the polynomial family we compute the $k^{\text{th}}$ moment for a large range of values of $k$, depending on the sizes of…
▽ More
We compute moments of $L$-functions associated to the polynomial, odd polynomial and ordinary families of Artin-Schreier covers over $\mathbb{F}_q$, where $q$ is a power of a prime $p$, when the size of the finite field is fixed and the genus of the family goes to infinity. In the polynomial family we compute the $k^{\text{th}}$ moment for a large range of values of $k$, depending on the sizes of $p$ and $q$. We also compute the second moment in absolute value of the polynomial family, obtaining an exact formula with a lower order term, and confirming the unitary symmetry type of the family. For the odd polynomial family, we obtain asymptotic formulas for the first two moments, in agreement with the symplectic random matrix theory model, and identifying a lower order term in the case of the first moment. We finally obtain an asymptotic formula for the first moment in the ordinary family of Artin-Schreier $L$-functions, again explicitly computing a lower order term.
△ Less
Submitted 28 June, 2023;
originally announced June 2023.
-
Negative moments of the Riemann zeta-function
Authors:
Hung M. Bui,
Alexandra Florea
Abstract:
Assuming the Riemann Hypothesis we study negative moments of the Riemann zeta-function and obtain asymptotic formulas in certain ranges of the shift in $ζ(s)$. For example, integrating $|ζ(1/2+α+it)|^{-2k}$ with respect to $t$ from $T$ to $2T$, we obtain an asymptotic formula when the shift $α$ is roughly bigger than $\frac{1}{\log T}$ and $k < 1/2$. We also obtain non-trivial upper bounds for muc…
▽ More
Assuming the Riemann Hypothesis we study negative moments of the Riemann zeta-function and obtain asymptotic formulas in certain ranges of the shift in $ζ(s)$. For example, integrating $|ζ(1/2+α+it)|^{-2k}$ with respect to $t$ from $T$ to $2T$, we obtain an asymptotic formula when the shift $α$ is roughly bigger than $\frac{1}{\log T}$ and $k < 1/2$. We also obtain non-trivial upper bounds for much smaller shifts, as long as $\log\frac{1}α \ll \log \log T$. This provides partial progress towards a conjecture of Gonek on negative moments of the Riemann zeta-function, and settles the conjecture in certain ranges. As an application, we also obtain an upper bound for the average of the generalized Möbius function.
△ Less
Submitted 14 February, 2023;
originally announced February 2023.
-
COVID-Net USPro: An Open-Source Explainable Few-Shot Deep Prototypical Network to Monitor and Detect COVID-19 Infection from Point-of-Care Ultrasound Images
Authors:
Jessy Song,
Ashkan Ebadi,
Adrian Florea,
Pengcheng Xi,
Stéphane Tremblay,
Alexander Wong
Abstract:
As the Coronavirus Disease 2019 (COVID-19) continues to impact many aspects of life and the global healthcare systems, the adoption of rapid and effective screening methods to prevent further spread of the virus and lessen the burden on healthcare providers is a necessity. As a cheap and widely accessible medical image modality, point-of-care ultrasound (POCUS) imaging allows radiologists to ident…
▽ More
As the Coronavirus Disease 2019 (COVID-19) continues to impact many aspects of life and the global healthcare systems, the adoption of rapid and effective screening methods to prevent further spread of the virus and lessen the burden on healthcare providers is a necessity. As a cheap and widely accessible medical image modality, point-of-care ultrasound (POCUS) imaging allows radiologists to identify symptoms and assess severity through visual inspection of the chest ultrasound images. Combined with the recent advancements in computer science, applications of deep learning techniques in medical image analysis have shown promising results, demonstrating that artificial intelligence-based solutions can accelerate the diagnosis of COVID-19 and lower the burden on healthcare professionals. However, the lack of a huge amount of well-annotated data poses a challenge in building effective deep neural networks in the case of novel diseases and pandemics. Motivated by this, we present COVID-Net USPro, an explainable few-shot deep prototypical network, that monitors and detects COVID-19 positive cases with high precision and recall from minimal ultrasound images. COVID-Net USPro achieves 99.65% overall accuracy, 99.7% recall and 99.67% precision for COVID-19 positive cases when trained with only 5 shots. The analytic pipeline and results were verified by our contributing clinician with extensive experience in POCUS interpretation, ensuring that the network makes decisions based on actual patterns.
△ Less
Submitted 4 January, 2023;
originally announced January 2023.
-
Power-saving error terms for the number of $D_4$-quartic extensions over a number field ordered by discriminant
Authors:
Alina Bucur,
Alexandra Florea,
Allechar Serrano López,
Ila Varma
Abstract:
We study the asymptotic count of dihedral quartic extensions over a fixed number field with bounded norm of the relative discriminant. The main term of this count (including a summation formula for the constant) can be found in the literature (see Cohen--Diaz y Diaz--Olivier for the statement without proof and see Klüners for a proof), but a power-saving for the error term has not been explicitly…
▽ More
We study the asymptotic count of dihedral quartic extensions over a fixed number field with bounded norm of the relative discriminant. The main term of this count (including a summation formula for the constant) can be found in the literature (see Cohen--Diaz y Diaz--Olivier for the statement without proof and see Klüners for a proof), but a power-saving for the error term has not been explicitly determined except in the case that the base field is $\mathbb{Q}$. In this article, we describe the argument for obtaining both the explicit main term and a power-saving error term for the number of $D_4$-quartic extensions over a general base number field ordered by the norms of their relative discriminants. We also give an extensive overview of the history and development of number field asymptotics.
△ Less
Submitted 27 September, 2022;
originally announced September 2022.
-
Tiered Pruning for Efficient Differentialble Inference-Aware Neural Architecture Search
Authors:
Sławomir Kierat,
Mateusz Sieniawski,
Denys Fridman,
Chen-Han Yu,
Szymon Migacz,
Paweł Morkisz,
Alex-Fit Florea
Abstract:
We propose three novel pruning techniques to improve the cost and results of inference-aware Differentiable Neural Architecture Search (DNAS). First, we introduce Prunode, a stochastic bi-path building block for DNAS, which can search over inner hidden dimensions with O(1) memory and compute complexity. Second, we present an algorithm for pruning blocks within a stochastic layer of the SuperNet du…
▽ More
We propose three novel pruning techniques to improve the cost and results of inference-aware Differentiable Neural Architecture Search (DNAS). First, we introduce Prunode, a stochastic bi-path building block for DNAS, which can search over inner hidden dimensions with O(1) memory and compute complexity. Second, we present an algorithm for pruning blocks within a stochastic layer of the SuperNet during the search. Third, we describe a novel technique for pruning unnecessary stochastic layers during the search. The optimized models resulting from the search are called PruNet and establishes a new state-of-the-art Pareto frontier for NVIDIA V100 in terms of inference latency for ImageNet Top-1 image classification accuracy. PruNet as a backbone also outperforms GPUNet and EfficientNet on the COCO object detection task on inference latency relative to mean Average Precision (mAP).
△ Less
Submitted 5 January, 2023; v1 submitted 23 September, 2022;
originally announced September 2022.
-
GPUNet: Searching the Deployable Convolution Neural Networks for GPUs
Authors:
Linnan Wang,
Chenhan Yu,
Satish Salian,
Slawomir Kierat,
Szymon Migacz,
Alex Fit Florea
Abstract:
Customizing Convolution Neural Networks (CNN) for production use has been a challenging task for DL practitioners. This paper intends to expedite the model customization with a model hub that contains the optimized models tiered by their inference latency using Neural Architecture Search (NAS). To achieve this goal, we build a distributed NAS system to search on a novel search space that consists…
▽ More
Customizing Convolution Neural Networks (CNN) for production use has been a challenging task for DL practitioners. This paper intends to expedite the model customization with a model hub that contains the optimized models tiered by their inference latency using Neural Architecture Search (NAS). To achieve this goal, we build a distributed NAS system to search on a novel search space that consists of prominent factors to impact latency and accuracy. Since we target GPU, we name the NAS optimized models as GPUNet, which establishes a new SOTA Pareto frontier in inference latency and accuracy. Within 1$ms$, GPUNet is 2x faster than EfficientNet-X and FBNetV3 with even better accuracy. We also validate GPUNet on detection tasks, and GPUNet consistently outperforms EfficientNet-X and FBNetV3 on COCO detection tasks in both latency and accuracy. All of these data validate that our NAS system is effective and generic to handle different design tasks. With this NAS system, we expand GPUNet to cover a wide range of latency targets such that DL practitioners can deploy our models directly in different scenarios.
△ Less
Submitted 26 April, 2022;
originally announced May 2022.
-
COVID-Net US-X: Enhanced Deep Neural Network for Detection of COVID-19 Patient Cases from Convex Ultrasound Imaging Through Extended Linear-Convex Ultrasound Augmentation Learning
Authors:
E. Zhixuan Zeng,
Adrian Florea,
Alexander Wong
Abstract:
As the global population continues to face significant negative impact by the on-going COVID-19 pandemic, there has been an increasing usage of point-of-care ultrasound (POCUS) imaging as a low-cost and effective imaging modality of choice in the COVID-19 clinical workflow. A major barrier with widespread adoption of POCUS in the COVID-19 clinical workflow is the scarcity of expert clinicians that…
▽ More
As the global population continues to face significant negative impact by the on-going COVID-19 pandemic, there has been an increasing usage of point-of-care ultrasound (POCUS) imaging as a low-cost and effective imaging modality of choice in the COVID-19 clinical workflow. A major barrier with widespread adoption of POCUS in the COVID-19 clinical workflow is the scarcity of expert clinicians that can interpret POCUS examinations, leading to considerable interest in deep learning-driven clinical decision support systems to tackle this challenge. A major challenge to building deep neural networks for COVID-19 screening using POCUS is the heterogeneity in the types of probes used to capture ultrasound images (e.g., convex vs. linear probes), which can lead to very different visual appearances. In this study, we explore the impact of leveraging extended linear-convex ultrasound augmentation learning on producing enhanced deep neural networks for COVID-19 assessment, where we conduct data augmentation on convex probe data alongside linear probe data that have been transformed to better resemble convex probe data. Experimental results using an efficient deep columnar anti-aliased convolutional neural network designed via a machined-driven design exploration strategy (which we name COVID-Net US-X) show that the proposed extended linear-convex ultrasound augmentation learning significantly increases performance, with a gain of 5.1% in test accuracy and 13.6% in AUC.
△ Less
Submitted 28 April, 2022;
originally announced April 2022.
-
COVID-Net Biochem: An Explainability-driven Framework to Building Machine Learning Models for Predicting Survival and Kidney Injury of COVID-19 Patients from Clinical and Biochemistry Data
Authors:
Hossein Aboutalebi,
Maya Pavlova,
Mohammad Javad Shafiee,
Adrian Florea,
Andrew Hryniowski,
Alexander Wong
Abstract:
Since the World Health Organization declared COVID-19 a pandemic in 2020, the global community has faced ongoing challenges in controlling and mitigating the transmission of the SARS-CoV-2 virus, as well as its evolving subvariants and recombinants. A significant challenge during the pandemic has not only been the accurate detection of positive cases but also the efficient prediction of risks asso…
▽ More
Since the World Health Organization declared COVID-19 a pandemic in 2020, the global community has faced ongoing challenges in controlling and mitigating the transmission of the SARS-CoV-2 virus, as well as its evolving subvariants and recombinants. A significant challenge during the pandemic has not only been the accurate detection of positive cases but also the efficient prediction of risks associated with complications and patient survival probabilities. These tasks entail considerable clinical resource allocation and attention.In this study, we introduce COVID-Net Biochem, a versatile and explainable framework for constructing machine learning models. We apply this framework to predict COVID-19 patient survival and the likelihood of develo** Acute Kidney Injury during hospitalization, utilizing clinical and biochemical data in a transparent, systematic approach. The proposed approach advances machine learning model design by seamlessly integrating domain expertise with explainability tools, enabling model decisions to be based on key biomarkers. This fosters a more transparent and interpretable decision-making process made by machines specifically for medical applications.
△ Less
Submitted 17 October, 2023; v1 submitted 24 April, 2022;
originally announced April 2022.
-
Negative moments of $L$-functions with small shifts over function fields
Authors:
Alexandra Florea
Abstract:
We consider negative moments of quadratic Dirichlet $L$--functions over function fields. Summing over monic square-free polynomials of degree $2g+1$ in $\mathbb{F}_q[x]$, we obtain an asymptotic formula for the $k^{\text{th}}$ shifted negative moment of $L(1/2+β,χ_D)$, in certain ranges of $β$ (for example, when roughly $β\gg \log g/g $ and $k<1$). We also obtain non-trivial upper bounds for the…
▽ More
We consider negative moments of quadratic Dirichlet $L$--functions over function fields. Summing over monic square-free polynomials of degree $2g+1$ in $\mathbb{F}_q[x]$, we obtain an asymptotic formula for the $k^{\text{th}}$ shifted negative moment of $L(1/2+β,χ_D)$, in certain ranges of $β$ (for example, when roughly $β\gg \log g/g $ and $k<1$). We also obtain non-trivial upper bounds for the $k^{\text{th}}$ shifted negative moment when $\log(1/β) \ll \log g$. Previously, almost sharp upper bounds were obtained in \cite{ratios} in the range $β\gg g^{-\frac{1}{2k}+ε}$.
△ Less
Submitted 28 November, 2022; v1 submitted 19 November, 2021;
originally announced November 2021.
-
The Ratios Conjecture and upper bounds for negative moments of $L$-functions over function fields
Authors:
Hung M. Bui,
Alexandra Florea,
Jonathan P. Keating
Abstract:
We prove special cases of the Ratios Conjecture for the family of quadratic Dirichlet $L$--functions over function fields. More specifically, we study the average of $L(1/2+α,χ_D)/L(1/2+β,χ_D)$, when $D$ varies over monic, square-free polynomials of degree $2g+1$ over $\mathbb{F}_q[x]$, as $g \to \infty$, and we obtain an asymptotic formula when $\Re β\gg g^{-1/2+\varepsilon}$. We also study avera…
▽ More
We prove special cases of the Ratios Conjecture for the family of quadratic Dirichlet $L$--functions over function fields. More specifically, we study the average of $L(1/2+α,χ_D)/L(1/2+β,χ_D)$, when $D$ varies over monic, square-free polynomials of degree $2g+1$ over $\mathbb{F}_q[x]$, as $g \to \infty$, and we obtain an asymptotic formula when $\Re β\gg g^{-1/2+\varepsilon}$. We also study averages of products of $2$ over $2$ and $3$ over $3$ $L$--functions, and obtain asymptotic formulas when the shifts in the denominator have real part bigger than $g^{-1/4+\varepsilon}$ and $g^{-1/6+\varepsilon}$ respectively. The main ingredient in the proof is obtaining upper bounds for negative moments of $L$--functions. The upper bounds we obtain are expected to be almost sharp in the ranges described above. As an application, we recover the asymptotic formula for the one-level density of zeros in the family with the support of the Fourier transform in $(-2,2)$.
△ Less
Submitted 21 September, 2021;
originally announced September 2021.
-
Hilbert transforms and the equidistribution of zeros of polynomials
Authors:
Emanuel Carneiro,
Mithun Kumar Das,
Alexandra Florea,
Angel V. Kumchev,
Amita Malik,
Micah B. Milinovich,
Caroline Turnage-Butterbaugh,
Jiuya Wang
Abstract:
We improve the current bounds for an inequality of Erdős and Turán from 1950 related to the discrepancy of angular equidistribution of the zeros of a given polynomial. Building upon a recent work of Soundararajan, we establish a novel connection between this inequality and an extremal problem in Fourier analysis involving the maxima of Hilbert transforms, for which we provide a complete solution.…
▽ More
We improve the current bounds for an inequality of Erdős and Turán from 1950 related to the discrepancy of angular equidistribution of the zeros of a given polynomial. Building upon a recent work of Soundararajan, we establish a novel connection between this inequality and an extremal problem in Fourier analysis involving the maxima of Hilbert transforms, for which we provide a complete solution. Prior to Soundararajan (2019), refinements of the discrepancy inequality of Erdős and Turán had been obtained by Ganelius (1954) and Mignotte (1992).
△ Less
Submitted 7 July, 2021; v1 submitted 31 March, 2021;
originally announced April 2021.
-
Iterative Knowledge Exchange Between Deep Learning and Space-Time Spectral Clustering for Unsupervised Segmentation in Videos
Authors:
Emanuela Haller,
Adina Magda Florea,
Marius Leordeanu
Abstract:
We propose a dual system for unsupervised object segmentation in video, which brings together two modules with complementary properties: a space-time graph that discovers objects in videos and a deep network that learns powerful object features. The system uses an iterative knowledge exchange policy. A novel spectral space-time clustering process on the graph produces unsupervised segmentation mas…
▽ More
We propose a dual system for unsupervised object segmentation in video, which brings together two modules with complementary properties: a space-time graph that discovers objects in videos and a deep network that learns powerful object features. The system uses an iterative knowledge exchange policy. A novel spectral space-time clustering process on the graph produces unsupervised segmentation masks passed to the network as pseudo-labels. The net learns to segment in single frames what the graph discovers in video and passes back to the graph strong image-level features that improve its node-level features in the next iteration. Knowledge is exchanged for several cycles until convergence. The graph has one node per each video pixel, but the object discovery is fast. It uses a novel power iteration algorithm computing the main space-time cluster as the principal eigenvector of a special Feature-Motion matrix without actually computing the matrix. The thorough experimental analysis validates our theoretical claims and proves the effectiveness of the cyclical knowledge exchange. We also perform experiments on the supervised scenario, incorporating features pretrained with human supervision. We achieve state-of-the-art level on unsupervised and supervised scenarios on four challenging datasets: DAVIS, SegTrack, YouTube-Objects, and DAVSOD.
△ Less
Submitted 13 December, 2020;
originally announced December 2020.
-
Pay Attention when Required
Authors:
Swetha Mandava,
Szymon Migacz,
Alex Fit Florea
Abstract:
Transformer-based models consist of interleaved feed-forward blocks - that capture content meaning, and relatively more expensive self-attention blocks - that capture context meaning. In this paper, we explored trade-offs and ordering of the blocks to improve upon the current Transformer architecture and proposed PAR Transformer. It needs 35% lower compute time than Transformer-XL achieved by repl…
▽ More
Transformer-based models consist of interleaved feed-forward blocks - that capture content meaning, and relatively more expensive self-attention blocks - that capture context meaning. In this paper, we explored trade-offs and ordering of the blocks to improve upon the current Transformer architecture and proposed PAR Transformer. It needs 35% lower compute time than Transformer-XL achieved by replacing ~63% of the self-attention blocks with feed-forward blocks, and retains the perplexity on WikiText-103 language modelling benchmark. We further validated our results on text8 and enwiki8 datasets, as well as on the BERT model.
△ Less
Submitted 17 May, 2021; v1 submitted 9 September, 2020;
originally announced September 2020.
-
Machine Learning Optimization of Quantum Circuit Layouts
Authors:
Alexandru Paler,
Lucian M. Sasu,
Adrian Florea,
Razvan Andonie
Abstract:
The quantum circuit layout (QCL) problem is to map a quantum circuit such that the constraints of the device are satisfied. We introduce a quantum circuit map** heuristic, QXX, and its machine learning version, QXX-MLP. The latter infers automatically the optimal QXX parameter values such that the layed out circuit has a reduced depth. In order to speed up circuit compilation, before laying the…
▽ More
The quantum circuit layout (QCL) problem is to map a quantum circuit such that the constraints of the device are satisfied. We introduce a quantum circuit map** heuristic, QXX, and its machine learning version, QXX-MLP. The latter infers automatically the optimal QXX parameter values such that the layed out circuit has a reduced depth. In order to speed up circuit compilation, before laying the circuits out, we are using a Gaussian function to estimate the depth of the compiled circuits. This Gaussian also informs the compiler about the circuit region that influences most the resulting circuit's depth. We present empiric evidence for the feasibility of learning the layout method using approximation. QXX and QXX-MLP open the path to feasible large scale QCL methods.
△ Less
Submitted 25 September, 2022; v1 submitted 29 July, 2020;
originally announced July 2020.
-
Non-vanishing for cubic $L$--functions
Authors:
Chantal David,
Alexandra Florea,
Matilde Lalin
Abstract:
We prove that there is a positive proportion of $L$-functions associated to cubic characters over $\mathbb{F}_q[T]$ that do not vanish at the critical point $s=1/2$. This is achieved by computing the first mollified moment using techniques previously developed by the authors in their work on the first moment of cubic $L$-functions, and by obtaining a sharp upper bound for the second mollified mome…
▽ More
We prove that there is a positive proportion of $L$-functions associated to cubic characters over $\mathbb{F}_q[T]$ that do not vanish at the critical point $s=1/2$. This is achieved by computing the first mollified moment using techniques previously developed by the authors in their work on the first moment of cubic $L$-functions, and by obtaining a sharp upper bound for the second mollified moment, building on work of Lester-Radziwill, which in turn develops further ideas from the work of Soundararajan, Harper, and Radziwill-Soundararajan. We work in the non-Kummer setting when $q \equiv 2\pmod{3}$, but our results could be translated into the Kummer setting when $q\equiv 1\pmod{3}$ as well as into the number field case (assuming the Generalized Riemann Hypothesis). Our positive proportion of non-vanishing is explicit, but extremely small, due to the fact that the implied constant in the upper bound for the mollified second moment is very large.
△ Less
Submitted 28 June, 2020;
originally announced June 2020.
-
Weighted Random Search for Hyperparameter Optimization
Authors:
Adrian-Catalin Florea,
Razvan Andonie
Abstract:
We introduce an improved version of Random Search (RS), used here for hyperparameter optimization of machine learning algorithms. Unlike the standard RS, which generates for each trial new values for all hyperparameters, we generate new values for each hyperparameter with a probability of change. The intuition behind our approach is that a value that already triggered a good result is a good candi…
▽ More
We introduce an improved version of Random Search (RS), used here for hyperparameter optimization of machine learning algorithms. Unlike the standard RS, which generates for each trial new values for all hyperparameters, we generate new values for each hyperparameter with a probability of change. The intuition behind our approach is that a value that already triggered a good result is a good candidate for the next step, and should be tested in new combinations of hyperparameter values. Within the same computational budget, our method yields better results than the standard RS. Our theoretical results prove this statement. We test our method on a variation of one of the most commonly used objective function for this class of problems (the Grievank function) and for the hyperparameter optimization of a deep learning CNN architecture. Our results can be generalized to any optimization problem defined on a discrete domain.
△ Less
Submitted 3 April, 2020;
originally announced April 2020.
-
Weighted Random Search for CNN Hyperparameter Optimization
Authors:
Razvan Andonie,
Adrian-Catalin Florea
Abstract:
Nearly all model algorithms used in machine learning use two different sets of parameters: the training parameters and the meta-parameters (hyperparameters). While the training parameters are learned during the training phase, the values of the hyperparameters have to be specified before learning starts. For a given dataset, we would like to find the optimal combination of hyperparameter values, i…
▽ More
Nearly all model algorithms used in machine learning use two different sets of parameters: the training parameters and the meta-parameters (hyperparameters). While the training parameters are learned during the training phase, the values of the hyperparameters have to be specified before learning starts. For a given dataset, we would like to find the optimal combination of hyperparameter values, in a reasonable amount of time. This is a challenging task because of its computational complexity. In previous work [11], we introduced the Weighted Random Search (WRS) method, a combination of Random Search (RS) and probabilistic greedy heuristic. In the current paper, we compare the WRS method with several state-of-the art hyperparameter optimization methods with respect to Convolutional Neural Network (CNN) hyperparameter optimization. The criterion is the classification accuracy achieved within the same number of tested combinations of hyperparameter values. According to our experiments, the WRS algorithm outperforms the other methods.
△ Less
Submitted 30 March, 2020;
originally announced March 2020.
-
Type-I contributions to the one and two level densities of quadratic Dirichlet $L$--functions over function fields
Authors:
Hung M. Bui,
Alexandra Florea,
Jonathan Keating
Abstract:
Using the Ratios Conjecture, we write down precise formulas with lower order terms for the one and the two level densities of zeros of quadratic Dirichlet $L$--functions over function fields. We denote the various terms arising as Type-$0$, Type-I and Type-II contributions. When the support of the Fourier transform of the test function is sufficiently restricted, we rigorously compute the Type-…
▽ More
Using the Ratios Conjecture, we write down precise formulas with lower order terms for the one and the two level densities of zeros of quadratic Dirichlet $L$--functions over function fields. We denote the various terms arising as Type-$0$, Type-I and Type-II contributions. When the support of the Fourier transform of the test function is sufficiently restricted, we rigorously compute the Type-$0$ and Type-I terms and confirm that they match the conjectured answer. When the restrictions on the support are relaxed, our results suggest that Type-II contributions become important in the two level density.
△ Less
Submitted 9 January, 2020;
originally announced January 2020.
-
Moments of Dirichlet $L$-functions with prime conductors over function fields
Authors:
Hung M. Bui,
Alexandra Florea
Abstract:
We compute the second moment in the family of quadratic Dirichlet $L$-functions with prime conductors over $\mathbb{F}_q[x]$ when the degree of the discriminant goes to infinity, obtaining one of the lower order terms. We also obtain an asymptotic formula with the leading order term for the mean value of the derivatives of $L$-functions associated to quadratic twists of a fixed elliptic curve over…
▽ More
We compute the second moment in the family of quadratic Dirichlet $L$-functions with prime conductors over $\mathbb{F}_q[x]$ when the degree of the discriminant goes to infinity, obtaining one of the lower order terms. We also obtain an asymptotic formula with the leading order term for the mean value of the derivatives of $L$-functions associated to quadratic twists of a fixed elliptic curve over $\mathbb{F}_q(t)$ by monic irreducible polynomials, which allows us to show that there exists a monic irreducible polynomial such that the analytic rank of the corresponding twisted elliptic curve is equal to $1$.
△ Less
Submitted 2 September, 2019;
originally announced September 2019.
-
Spacetime Graph Optimization for Video Object Segmentation
Authors:
Emanuela Haller,
Adina Magda Florea,
Marius Leordeanu
Abstract:
We address the challenging task of foreground object discovery and segmentation in video. We introduce an efficient solution, suitable for both unsupervised and supervised scenarios, based on a spacetime graph representation of the video sequence. We ensure a fine grained representation with one-to-one correspondences between graph nodes and video pixels. We formulate the task as a spectral cluste…
▽ More
We address the challenging task of foreground object discovery and segmentation in video. We introduce an efficient solution, suitable for both unsupervised and supervised scenarios, based on a spacetime graph representation of the video sequence. We ensure a fine grained representation with one-to-one correspondences between graph nodes and video pixels. We formulate the task as a spectral clustering problem by exploiting the spatio-temporal consistency between the scene elements in terms of motion and appearance. Graph nodes that belong to the main object of interest should form a strong cluster, as they are linked through long range optical flow chains and have similar motion and appearance features along those chains. On one hand, the optimization problem aims to maximize the segmentation clustering score based on the motion structure through space and time. On the other hand, the segmentation should be consistent with respect to node features. Our approach leads to a graph formulation in which the segmentation solution becomes the principal eigenvector of a novel Feature-Motion matrix. While the actual matrix is not computed explicitly, the proposed algorithm efficiently computes, in a few iteration steps, the principal eigenvector that captures the segmentation of the main object in the video. The proposed algorithm, GO-VOS, produces a global optimum solution and, consequently, it does not depend on initialization. In practice, GO-VOS achieves state of the art results on three challenging datasets used in current literature: DAVIS, SegTrack and YouTube-Objects.
△ Less
Submitted 3 August, 2019; v1 submitted 7 July, 2019;
originally announced July 2019.
-
Moments of quadratic twists of elliptic curve L-functions over function fields
Authors:
H. M. Bui,
Alexandra Florea,
Jonathan P. Keating,
Edva Roditty-Gershon
Abstract:
We calculate the first and second moments of L-functions in the family of quadratic twists of a fixed elliptic curve E over F_q[x], asymptotically in the limit as the degree of the twists tends to infinity. We also compute moments involving derivatives of L-functions over quadratic twists, enabling us to deduce lower bounds on the correlations between the analytic ranks of the twists of two distin…
▽ More
We calculate the first and second moments of L-functions in the family of quadratic twists of a fixed elliptic curve E over F_q[x], asymptotically in the limit as the degree of the twists tends to infinity. We also compute moments involving derivatives of L-functions over quadratic twists, enabling us to deduce lower bounds on the correlations between the analytic ranks of the twists of two distinct curves.
△ Less
Submitted 1 February, 2019;
originally announced February 2019.
-
The mean values of cubic L-functions over function fields
Authors:
Chantal David,
Alexandra Florea,
Matilde Lalin
Abstract:
We obtain an asymptotic formula for the mean value of L-functions associated to cubic characters over F_q[t]. We solve this problem in the non-Kummer setting when q=2 (mod 3) and in the Kummer case when q=1 (mod 3). The proofs rely on obtaining precise asymptotics for averages of cubic Gauss sums over function fields, which can be studied using the theory of metaplectic Eisenstein series. In the n…
▽ More
We obtain an asymptotic formula for the mean value of L-functions associated to cubic characters over F_q[t]. We solve this problem in the non-Kummer setting when q=2 (mod 3) and in the Kummer case when q=1 (mod 3). The proofs rely on obtaining precise asymptotics for averages of cubic Gauss sums over function fields, which can be studied using the theory of metaplectic Eisenstein series. In the non-Kummer setting we display some explicit cancellation between the main term and the dual term coming from the approximate functional equation of the L-functions.
△ Less
Submitted 3 January, 2019;
originally announced January 2019.
-
Hybrid Euler-Hadamard product for quadratic Dirichlet $L$-functions in function fields
Authors:
H. M. Bui,
Alexandra Florea
Abstract:
We develop a hybrid Euler-Hadamard product model for quadratic Dirichlet $L$--functions over function fields (following the model introduced by Gonek, Hughes and Keating for the Riemann-zeta function). After computing the first three twisted moments in this family of $L$--functions, we provide further evidence for the conjectural asymptotic formulas for the moments of the family.
We develop a hybrid Euler-Hadamard product model for quadratic Dirichlet $L$--functions over function fields (following the model introduced by Gonek, Hughes and Keating for the Riemann-zeta function). After computing the first three twisted moments in this family of $L$--functions, we provide further evidence for the conjectural asymptotic formulas for the moments of the family.
△ Less
Submitted 17 September, 2016;
originally announced September 2016.
-
The fourth moment of quadratic Dirichlet $L$--functions over function fields
Authors:
Alexandra Florea
Abstract:
We obtain an asymptotic formula for the fourth moment of quadratic Dirichlet $L$--functions over $\mathbb{F}_q[x]$, as the base field $\mathbb{F}_q$ is fixed and the genus of the family goes to infinity. According to conjectures of Andrade and Keating, we expect the fourth moment to be asymptotic to $q^{2g+1} P(2g+1)$ up to an error of size $o(q^{2g+1})$, where $P$ is a polynomial of degree $10$ w…
▽ More
We obtain an asymptotic formula for the fourth moment of quadratic Dirichlet $L$--functions over $\mathbb{F}_q[x]$, as the base field $\mathbb{F}_q$ is fixed and the genus of the family goes to infinity. According to conjectures of Andrade and Keating, we expect the fourth moment to be asymptotic to $q^{2g+1} P(2g+1)$ up to an error of size $o(q^{2g+1})$, where $P$ is a polynomial of degree $10$ with explicit coefficients. We prove an asymptotic formula with the leading three terms, which agrees with the conjectured result.
△ Less
Submitted 5 September, 2016;
originally announced September 2016.
-
Zeros of quadratic Dirichlet $L$-functions in the hyperelliptic ensemble
Authors:
Hung M. Bui,
Alexandra Florea
Abstract:
We study the $1$-level density and the pair correlation of zeros of quadratic Dirichlet $L$-functions in function fields, as we average over the ensemble $\mathcal{H}_{2g+1}$ of monic, square-free polynomials with coefficients in $\mathbb{F}_q[x]$. In the case of the $1$-level density, when the Fourier transform of the test function is supported in the restricted interval $(\frac{1}{3},1)$, we com…
▽ More
We study the $1$-level density and the pair correlation of zeros of quadratic Dirichlet $L$-functions in function fields, as we average over the ensemble $\mathcal{H}_{2g+1}$ of monic, square-free polynomials with coefficients in $\mathbb{F}_q[x]$. In the case of the $1$-level density, when the Fourier transform of the test function is supported in the restricted interval $(\frac{1}{3},1)$, we compute a secondary term of size $q^{-\frac{4g}{3}}/g$, which is not predicted by the Ratios Conjecture. Moreover, when the support is even more restricted, we obtain several lower order terms. For example, if the Fourier transform is supported in $(\frac{1}{3}, \frac{1}{2})$, we identify another lower order term of size $q^{-\frac{8g}{5}}/g$. We also compute the pair correlation, and as for the $1$-level density, we detect lower order terms under certain restrictions; for example, we see a term of size $q^{-g}/g^2$ when the Fourier transform is supported in $(\frac{1}{4},\frac{1}{2})$. The $1$-level density and the pair correlation allow us to obtain non-vanishing results for $L(\frac12,χ_D)$, as well as lower bounds for the proportion of simple zeros of this family of $L$-functions.
△ Less
Submitted 23 May, 2016;
originally announced May 2016.
-
The second and third moment of $L(\frac{1}{2},χ)$ in the hyperelliptic ensemble
Authors:
Alexandra Florea
Abstract:
We obtain asymptotic formulas for the second and third moment of quadratic Dirichlet $L$--functions at the critical point, in the function field setting. We fix the ground field $\mathbb{F}_q$, and assume for simplicity that $q$ is a prime with $q \equiv 1 \pmod 4$. We compute the second and third moment of $L(1/2,χ_D)$ when $D$ is a monic, square-free polynomial of degree $2g+1$, as…
▽ More
We obtain asymptotic formulas for the second and third moment of quadratic Dirichlet $L$--functions at the critical point, in the function field setting. We fix the ground field $\mathbb{F}_q$, and assume for simplicity that $q$ is a prime with $q \equiv 1 \pmod 4$. We compute the second and third moment of $L(1/2,χ_D)$ when $D$ is a monic, square-free polynomial of degree $2g+1$, as $g \to \infty$. The answer we get for the second moment agrees with Andrade and Keating's conjectured formula. For the third moment, we check that the leading term agrees with the conjecture.
△ Less
Submitted 9 July, 2015;
originally announced July 2015.
-
Improving the error term in the mean value of $L(\tfrac{1}{2},χ)$ in the hyperelliptic ensemble
Authors:
Alexandra Florea
Abstract:
Andrade and Keating computed the mean value of quadratic Dirichlet $L$--functions at the critical point, in the hyperelliptic ensemble over a fixed finite field $\mathbb{F}_q$. Summing $L(1/2,χ_D)$ over monic, square-free polynomials $D$ of degree $2g+1$, the main term is of size $|D| \log_q |D|$ (where $|D|=q^{2g+1}$) and Andrade and Keating bound the error term by…
▽ More
Andrade and Keating computed the mean value of quadratic Dirichlet $L$--functions at the critical point, in the hyperelliptic ensemble over a fixed finite field $\mathbb{F}_q$. Summing $L(1/2,χ_D)$ over monic, square-free polynomials $D$ of degree $2g+1$, the main term is of size $|D| \log_q |D|$ (where $|D|=q^{2g+1}$) and Andrade and Keating bound the error term by $|D|^{\frac 34+ \frac{\log_q(2)}{2}}$. For simplicity, we assume that $q$ is prime with $q \equiv 1 \pmod 4$. We prove that there is an extra term of size $|D|^{1/3} \log_q|D|$ in the asymptotic formula and bound the error term by $|D|^{1/4+ε}$.
△ Less
Submitted 12 May, 2015;
originally announced May 2015.
