Search

In computational digital pathology, accurate nuclear segmentation of Hematoxylin and Eosin (H&E) stained whole slide images (WSIs) is a critical step for many analyses and tissue characterizations. One popular deep learning-based nuclear segmentation approach, HoverNet, offers remarkably accurate results but lacks the high-throughput performance needed for clinical deployment in resource-constrained settings. Our approach, HoverFast, aims to provide fast and accurate nuclear segmentation in H&E images using knowledge distillation from HoverNet. By redesigning the tool with software engineering best practices, HoverFast introduces advanced parallel processing capabilities, efficient data handling, and optimized postprocessing. These improvements facilitate scalable high-throughput performance, making HoverFast more suitable for real-time analysis and application in resource-limited environments. Using a consumer grade Nvidia A5000 GPU, HoverFast showed a 21x speed improvement as compared to HoverNet; reducing mean analysis time for 40x WSIs from ~2 hours to 6 minutes while retaining a concordant mean Dice score of 0.91 against the original HoverNet output. Peak memory usage was also reduced 71% from 44.4GB, to 12.8GB, without requiring SSD-based caching. To ease adoption in research and clinical contexts, HoverFast aligns with best-practices in terms of (a) installation, and (b) containerization, while (c) providing outputs compatible with existing popular open-source image viewing tools such as QuPath. HoverFast has been made open-source and is available at andrewjanowczyk.com/open-source-tools/hoverfast., Comment: 9 pages, 3 figures, 1 appendix

We provide general estimates which compare the quermassintegrals of a convex body $K$ in ${\mathbb R}^n$ with the averages of the corresponding quermassintegrals of the $k$-codimensional sections of $K$ over $G_{n,n-k}$. An example is the inequality $$\alpha_{n,k,j}\frac{W_j(K)}{|K|}\leq\int_{G_{n,n-k}}\frac{W_j(K\cap F)}{|K\cap F|}d\nu_{n,n-k}(F)\leq \beta_{n,k,j}\frac{W_j(K)}{|K|}$$ where the constants $\alpha_{n,k,j}$ and $\beta_{n,k,j}$ depend only on $n,k$ and $j$, which holds true for any centrally symmetric convex body $K$ in ${\mathbb R}^n$ and any $0\leq j\leq n-k-1\leq n-1$. Using these estimates we obtain some positive results for suitable versions of the slicing problem for the quermassintegrals of a convex body.

In decentralized optimization environments, each agent $i$ in a network of $n$ nodes has its own private function $f_i$, and nodes communicate with their neighbors to cooperatively minimize the aggregate objective $\sum_{i=1}^n f_i$. In this setting, synchronizing the nodes' updates incurs significant communication overhead and computational costs, so much of the recent literature has focused on the analysis and design of asynchronous optimization algorithms, where agents activate and communicate at arbitrary times without needing a global synchronization enforcer. However, most works assume that when a node activates, it selects the neighbor to contact based on a fixed probability (e.g., uniformly at random), a choice that ignores the optimization landscape at the moment of activation. Instead, in this work we introduce an optimization-aware selection rule that chooses the neighbor providing the highest dual cost improvement (a quantity related to a dualization of the problem based on consensus). This scheme is related to the coordinate descent (CD) method with the Gauss-Southwell (GS) rule for coordinate updates; in our setting however, only a subset of coordinates is accessible at each iteration (because each node can communicate only with its neighbors), so the existing literature on GS methods does not apply. To overcome this difficulty, we develop a new analytical framework for smooth and strongly convex $f_i$ that covers the class of set-wise CD algorithms -- a class that directly applies to decentralized scenarios, but is not limited to them -- and we show that the proposed set-wise GS rule achieves a speedup factor of up to the maximum degree in the network (which is in the order of $\Theta(n)$ for highly connected graphs). The speedup predicted by our analysis is validated in numerical experiments with synthetic data., Comment: Revised writing, added references

Purpose - To develop and validate a deep learning (DL) framework for the detection and quantification of drusen and reticular pseudodrusen (RPD) on optical coherence tomography scans. Design - Development and validation of deep learning models for classification and feature segmentation. Methods - A DL framework was developed consisting of a classification model and an out-of-distribution (OOD) detection model for the identification of ungradable scans; a classification model to identify scans with drusen or RPD; and an image segmentation model to independently segment lesions as RPD or drusen. Data were obtained from 1284 participants in the UK Biobank (UKBB) with a self-reported diagnosis of age-related macular degeneration (AMD) and 250 UKBB controls. Drusen and RPD were manually delineated by five retina specialists. The main outcome measures were sensitivity, specificity, area under the ROC curve (AUC), kappa, accuracy and intraclass correlation coefficient (ICC). Results - The classification models performed strongly at their respective tasks (0.95, 0.93, and 0.99 AUC, respectively, for the ungradable scans classifier, the OOD model, and the drusen and RPD classification model). The mean ICC for drusen and RPD area vs. graders was 0.74 and 0.61, respectively, compared with 0.69 and 0.68 for intergrader agreement. FROC curves showed that the model's sensitivity was close to human performance. Conclusions - The models achieved high classification and segmentation performance, similar to human performance. Application of this robust framework will further our understanding of RPD as a separate entity from drusen in both research and clinical settings., Comment: 26 pages, 7 figures

We study the slicing inequality for the surface area instead of volume. This is the question whether there exists a constant $\alpha_n$ depending (or not) on the dimension $n$ so that $$S(K)\leq\alpha_n|K|^{\frac{1}{n}}\max_{\xi\in S^{n-1}}S(K\cap\xi^{\perp })$$ where $S$ denotes surface area and $|\cdot |$ denotes volume. For any fixed dimension we provide a negative answer to this question, as well as to a weaker version in which sections are replaced by projections onto hyperplanes. We also study the same problem for sections and projections of lower dimension and for all the quermassintegrals of a convex body. Starting from these questions, we also introduce a number of natural parameters relating volume and surface area, and provide optimal upper and lower bounds for them. Finally, we show that, in contrast to the previous negative results, a variant of the problem which arises naturally from the surface area version of the equivalence of the isomorphic Busemann--Petty problem with the slicing problem has an affirmative answer.

In this paper, we propose a novel algorithm for video rate adaptation in HTTP Adaptive Streaming (HAS), based on online learning. The proposed algorithm, named Learn2Adapt (L2A), is shown to provide a robust rate adaptation strategy which, unlike most of the state-of-the-art techniques, does not require parameter tuning, channel model assumptions or application-specific adjustments. These properties make it very suitable for mobile users, who typically experience fast variations in channel characteristics. Simulations show that L2A improves on the overall Quality of Experience (QoE) and in particular the average streaming rate, a result obtained independently of the channel and application scenarios., Comment: 9 pages, 3 figures, submitted to IEEE Transactions on Multimedia (under review)

Future 5G wireless networks will rely on agile and automated network management, where the usage of diverse resources must be jointly optimized with surgical accuracy. A number of key wireless network functionalities (e.g., traffic steering, power control) give rise to hard optimization problems. What is more, high spatio-temporal traffic variability coupled with the need to satisfy strict per slice/service SLAs in modern networks, suggest that these problems must be constantly (re-)solved, to maintain close-to-optimal performance. To this end, we propose the framework of Online Network Optimization (ONO), which seeks to maintain both agile and efficient control over time, using an arsenal of data-driven, online learning, and AI-based techniques. Since the mathematical tools and the studied regimes vary widely among these methodologies, a theoretical comparison is often out of reach. Therefore, the important question `what is the right ONO technique?' remains open to date. In this paper, we discuss the pros and cons of each technique and present a direct quantitative comparison for a specific use case, using real data. Our results suggest that carefully combining the insights of problem modeling with state-of-the-art AI techniques provides significant advantages at reasonable complexity.

The massive scale of future wireless networks will create computational bottlenecks in performance optimization. In this paper, we study the problem of connecting mobile traffic to Cloud RAN (C-RAN) stations. To balance station load, we steer the traffic by designing device association rules. The baseline association rule connects each device to the station with the strongest signal, which does not account for interference or traffic hot spots, and leads to load imbalances and performance deterioration. Instead, we can formulate an optimization problem to decide centrally the best association rule at each time instance. However, in practice this optimization has such high dimensions, that even linear programming solvers fail to solve. To address the challenge of massive connectivity, we propose an approach based on the theory of optimal transport, which studies the economical transfer of probability between two distributions. Our proposed methodology can further inspire scalable algorithms for massive optimization problems in wireless networks., Comment: 6 pages, 5 figures

The purpose of this article is to establish the dual version of the uniform cover inequality of Bollobas and Thomason.

The classical Loomis-Whitney inequality and the uniform cover inequality of Bollob\'{a}s and Thomason provide lower bounds for the volume of a compact set in terms of its lower dimensional coordinate projections. We provide further extensions of these inequalities in the setting of convex bodies. We also establish the corresponding dual inequalities for coordinate sections; these uniform cover inequalities for sections may be viewed as extensions of Meyer's dual Loomis-Whitney inequality.

We present an alternative approach to some results of Koldobsky on measures of sections of symmetric convex bodies, which allows us to extend them to the not necessarily symmetric setting. We prove that if $K$ is a convex body in ${\mathbb R}^n$ with $0\in {\rm int}(K)$ and if $\mu $ is a measure on ${\mathbb R}^n$ with a locally integrable non-negative density $g$ on ${\mathbb R}^n$, then \begin{equation*}\mu (K)\leq \left (c\sqrt{n-k}\right )^k\max_{F\in G_{n,n-k}}\mu (K\cap F)\cdot |K|^{\frac{k}{n}}\end{equation*} for every $1\leq k\leq n-1$. Also, if $\mu $ is even and log-concave, and if $K$ is a symmetric convex body in ${\mathbb R}^n$ and $D$ is a compact subset of ${\mathbb R}^n$ such that $\mu (K\cap F)\leq \mu (D\cap F)$ for all $F\in G_{n,n-k}$, then \begin{equation*}\mu (K)\leq \left (ckL_{n-k}\right )^{k}\mu (D),\end{equation*} where $L_s$ is the maximal isotropic constant of a convex body in ${\mathbb R}^s$. Our method employs a generalized Blaschke-Petkantschin formula and estimates for the dual affine quermassintegrals.

Creating and monitoring competitive and cost-effective pay-per-click advertisement campaigns through the web-search channel is a resource demanding task in terms of expertise and effort. Assisting or even automating the work of an advertising specialist will have an unrivaled commercial value. In this paper we propose a methodology, an architecture, and a fully functional framework for semi- and fully- automated creation, monitoring, and optimization of cost-efficient pay-per-click campaigns with budget constraints. The campaign creation module generates automatically keywords based on the content of the web page to be advertised extended with corresponding ad-texts. These keywords are used to create automatically the campaigns fully equipped with the appropriate values set. The campaigns are uploaded to the auctioneer platform and start running. The optimization module focuses on the learning process from existing campaign statistics and also from applied strategies of previous periods in order to invest optimally in the next period. The objective is to maximize the performance (i.e. clicks, actions) under the current budget constraint. The fully functional prototype is experimentally evaluated on real world Google AdWords campaigns and presents a promising behavior with regards to campaign performance statistics as it outperforms systematically the competing manually maintained campaigns., Comment: Submitted to ACM Transactions on the Web (TWEB) Journal - July 2012

A low-dimensional dynamic model for transitional buoyancy-driven flow in a differentially heated tall enclosure is presented. The full governing partial differential equations with the associated boundary conditions are solved by a spectral element method for a cavity of aspect ratio A=20. Proper orthogonal decomposition is applied to the oscillatory solution at Prandtl number Pr=P tau (omega) = 0.71 and Grashof number G tau (omega) = 3.2 x 10 (exp 4) to construct empirical eigenfunctions. Using the four most energetic empirical eigenfunctions for the velocity and temperature as basis functions and applying Galerkin's method, a reduced model consisting of eight nonlinear ordinary differential equations is obtained. Close to the 'design' conditions (P tau(omega) G tau(omega)), the low-order model (LOM) predictions are in excellent agreement with the predictions of the full model. In particular, the critical Grashof number at the onset of the first temporal flow instability (Hopf bifurcation) was well as the frequency and amplitude of oscillations at supercritical conditions are in excellent agreement with the predictions of the full model. Far from the 'design' conditions, the LOM predicts the existence of multiple stable steady solutions at large values of G tau, and a unique stable steady solution at small values of G tau, and exhibits hysteretic behavior that is qualitatively similar to that observed in direct numerical simulations based on the full model.