Search

How can we test AI performance? This question seems trivial, but it isn't. Standard benchmarks often have problems such as in-distribution and small-size test sets, oversimplified metrics, unfair comparisons, and short-term outcome pressure. As a consequence, good performance on standard benchmarks does not guarantee success in real-world scenarios. To address these problems, we present Touchstone, a large-scale collaborative segmentation benchmark of 9 types of abdominal organs. This benchmark is based on 5,195 training CT scans from 76 hospitals around the world and 5,903 testing CT scans from 11 additional hospitals. This diverse test set enhances the statistical significance of benchmark results and rigorously evaluates AI algorithms across various out-of-distribution scenarios. We invited 14 inventors of 19 AI algorithms to train their algorithms, while our team, as a third party, independently evaluated these algorithms on three test sets. In addition, we also evaluated pre-existing AI frameworks--which, differing from algorithms, are more flexible and can support different algorithms--including MONAI from NVIDIA, nnU-Net from DKFZ, and numerous other open-source frameworks. We are committed to expanding this benchmark to encourage more innovation of AI algorithms for the medical domain., Comment: Accepted to NeurIPS-2024

A primary challenge in large language model (LLM) development is their onerous pre-training cost. Typically, such pre-training involves optimizing a self-supervised objective (such as next-token prediction) over a large corpus. This paper explores a promising paradigm to improve LLM pre-training efficiency and quality by suitably leveraging a small language model (SLM). In particular, this paradigm relies on an SLM to both (1) provide soft labels as additional training supervision, and (2) select a small subset of valuable ("informative" and "hard") training examples. Put together, this enables an effective transfer of the SLM's predictive distribution to the LLM, while prioritizing specific regions of the training data distribution. Empirically, this leads to reduced LLM training time compared to standard training, while improving the overall quality. Theoretically, we develop a statistical framework to systematically study the utility of SLMs in enabling efficient training of high-quality LLMs. In particular, our framework characterizes how the SLM's seemingly low-quality supervision can enhance the training of a much more capable LLM. Furthermore, it also highlights the need for an adaptive utilization of such supervision, by striking a balance between the bias and variance introduced by the SLM-provided soft labels. We corroborate our theoretical framework by improving the pre-training of an LLM with 2.8B parameters by utilizing a smaller LM with 1.5B parameters on the Pile dataset.

We present a novel soft prompt based framework, SoftSRV, that leverages a frozen pre-trained large language model (LLM) to generate targeted synthetic text sequences. Given a sample from the target distribution, our proposed framework uses data-driven loss minimization to train a parameterized "contextual" soft prompt. This soft prompt is then used to steer the frozen LLM to generate synthetic sequences that are similar to the target distribution. We argue that SoftSRV provides a practical improvement over common hard-prompting approaches that rely on human-curated prompt-templates, which can be idiosyncratic, labor-intensive to craft, and may need to be specialized per domain. We empirically evaluate SoftSRV and hard-prompting baselines by generating synthetic data to fine-tune a small Gemma model on three different domains (coding, math, reasoning). To stress the generality of SoftSRV, we perform these evaluations without any particular specialization of the framework to each domain. We find that SoftSRV significantly improves upon hard-prompting baselines, generating data with superior fine-tuning performance and that better matches the target distribution according to the MAUVE similarity metric.

In recent years, understanding thermodynamics in the quantum regime has garnered significant attention, driven by advances in nanoscale physics and experimental techniques. In parallel, growing evidence supports the importance of quantum effects in various biological processes, making them increasingly relevant to quantum thermodynamics. In this study, we apply resource theory formulations of thermodynamics to investigate the role of quantum properties in ion transport across cell membranes. Within this framework, quantum properties are treated as resources under generalized thermodynamic constraints in the quantum regime. Specifically, our findings reveal that non-Markovianity, which reflects memory effects in ion transport dynamics, serves as a key quantum resource that enhances the yield and efficiency of the ion transport process. In contrast, quantum coherence, manifested as the superposition of energy states in ion-transport proteins, reduces these metrics but plays a crucial role in distinguishing between ion channels and ion pumps: two distinct types of ion-transport proteins in cell membranes. Finally, we demonstrate that introducing an additional coherent system allows coherence to facilitate the transformation of an ion pump into an ion channel., Comment: 9 pages, 3 figures

We present MM1.5, a new family of multimodal large language models (MLLMs) designed to enhance capabilities in text-rich image understanding, visual referring and grounding, and multi-image reasoning. Building upon the MM1 architecture, MM1.5 adopts a data-centric approach to model training, systematically exploring the impact of diverse data mixtures across the entire model training lifecycle. This includes high-quality OCR data and synthetic captions for continual pre-training, as well as an optimized visual instruction-tuning data mixture for supervised fine-tuning. Our models range from 1B to 30B parameters, encompassing both dense and mixture-of-experts (MoE) variants, and demonstrate that careful data curation and training strategies can yield strong performance even at small scales (1B and 3B). Additionally, we introduce two specialized variants: MM1.5-Video, designed for video understanding, and MM1.5-UI, tailored for mobile UI understanding. Through extensive empirical studies and ablations, we provide detailed insights into the training processes and decisions that inform our final designs, offering valuable guidance for future research in MLLM development.

We investigate experimentally the undirected open microwave network $\Gamma $ with internal absorption composed of two coupled directed halves, unidirectional networks $\Gamma_{+} $ and $\Gamma_{-} $, corresponding to two possible directions of motion on their edges. The two-port scattering matrix of the network $\Gamma$ is measured and the spectral statistics and the elastic enhancement factor of the network are evaluated. The comparison of the number of experimental resonances with the theoretical one predicted by the Weyl's law shows that within the experimental resolution the resonances are doubly degenerate. This conclusion was also corroborated by the numerical calculations. Though the network is characterized by the time reversal symmetry the missing level spectral statistics and the elastic enhancement factor are rather close to the Gaussian unitary ensemble predictions in random matrix theory. We used numerical calculations for the open non-dissipative quantum graph possessing the same structure as the microwave network $\Gamma$ to investigate the doublet structures in the spectrum which otherwise would not be experimentally resolved. We show that the doublet size distribution is close to the Poisson distribution.

Fog computing is essentially the expansion of cloud computing towards the network edge, reducing user access time to computing resources and services. Various advantages attribute to fog computing, including reduced latency, and improved user experience. However, user mobility may limit the benefits of fog computing. The displacement of users from one location to another, may increase their distance from a fog server, leading into latency amplification. This would also increase the probability of over utilization of fog servers which are located in popular destinations of mobile edge devices. This creates an unbalanced network of fog devices failing to provide lower makespan and fewer cloud accesses. One solution to maintain latency within an acceptable range is the migration of fog tasks and preserve the distance between the edge devices and the available resources. Although some studies have focused on fog task migration, none of them have considered load balancing in fog nodes. Accordingly, this paper introduces LIMO; an allocation and migration strategy for establishing load balancing in fog networks based on the control loop MAPE (Monitor-Analyze-Plan-Execute) and the Particle Swarm Optimization (PSO) algorithm. The periodical migration of tasks for load balancing aims to enhance the system's efficiency. The performance of LIMO has been modeled and evaluated using the Mobfogsim toolkit. The results show that this technique outperforms the state-of-the-art in terms of network resource utilization with 10% improvement. Furthermore, LIMO reduces the task migration to cloud by more than 15%, while it reduces the request response time by 18%., Comment: This paper has been accepted in CPSAT 2024 and will be published in the IEEE Xplore Digital Library

A meta-analysis of spaceflight data from both mouse and human flights reveals a striking overlap with Parkinson's disease (PD). Parallels include: changes in gait, loss of dopamine, sustained changes in the basal ganglia, loss of tyrosine hydroxylase in the substantia nigra, and systemic mitochondrial dysfunction. We identified specific Parkinson's genes differentially expressed post-spaceflight. These evidences indicate that spaceflight stressor-induced changes in the brain may become permanent during deep space exploration, posing a risk of PD in astronauts, Comment: 2 figures and 1 table

A generalized dispersion equation is derived featuring coupled mode theory, parity-time symmetry, and leaky wave antennas of arbitrary periodic modulation. It can be specialized to each of these cases individually or can describe a structure containing all three electromagnetic phenomena simultaneously in a single antenna. This very general dispersion equation is derived using both mode matching and the transverse resonance method, the latter lacking the ability to provide the field descriptions and wave impedances at the cost of computational simplicity. Using the dispersion equation, a sinusoidally modulated reactive sheet (SMRS) supported by an active impedance sheet backed dielectric spacer is designed. The active impedance sheet is designed to compensate the SMRS radiative leakage loss when coupled in close proximity. Due to the coupling, each spatial harmonic generated by the SMRS modulation is characterized by a purely real propagation constant and hence leaky wave radiation is generated from surface waves which do not decay. Due to the non-decaying nature of the surface waves, an input to output surface wave port-to-port |S21| would be unity despite an open far field channel being generated. The non-decaying leaky surface wave antenna is compared to the traditional PEC backed SMRS. The non-decaying leaky surface waves give rise to perfect aperture efficiency and the associated radiation pattern making it far superior to the traditional PEC backed SMRS case. Full-wave simulations corroborate the results. These types of antennas can make sensors probed from the far field where changes in S21 amplitude directly correlate to local changes in the waveguiding environment.

This paper proposes a hybrid fusion-based deep learning approach based on two different modalities, audio and video, to improve human activity recognition and violence detection in public places. To take advantage of audiovisual fusion, late fusion, intermediate fusion, and hybrid fusion-based deep learning (HFBDL) are used and compared. Since the objective is to detect and recognize human violence in public places, Real-life violence situation (RLVS) dataset is expanded and used. Simulating results of HFBDL show 96.67\% accuracy on validation data, which is more accurate than the other state-of-the-art methods on this dataset. To showcase our model's ability in real-world scenarios, another dataset of 54 sounded videos of both violent and non-violent situations was recorded. The model could successfully detect 52 out of 54 videos correctly. The proposed method shows a promising performance on real scenarios. Thus, it can be used for human action recognition and violence detection in public places for security purposes., Comment: This work has been submitted to the IEEE for possible publication, 10 pages, 8 figures

Medical image segmentation involves identifying and separating object instances in a medical image to delineate various tissues and structures, a task complicated by the significant variations in size, shape, and density of these features. Convolutional neural networks (CNNs) have traditionally been used for this task but have limitations in capturing long-range dependencies. Transformers, equipped with self-attention mechanisms, aim to address this problem. However, in medical image segmentation it is beneficial to merge both local and global features to effectively integrate feature maps across various scales, capturing both detailed features and broader semantic elements for dealing with variations in structures. In this paper, we introduce MSA$^2$Net, a new deep segmentation framework featuring an expedient design of skip-connections. These connections facilitate feature fusion by dynamically weighting and combining coarse-grained encoder features with fine-grained decoder feature maps. Specifically, we propose a Multi-Scale Adaptive Spatial Attention Gate (MASAG), which dynamically adjusts the receptive field (Local and Global contextual information) to ensure that spatially relevant features are selectively highlighted while minimizing background distractions. Extensive evaluations involving dermatology, and radiological datasets demonstrate that our MSA$^2$Net outperforms state-of-the-art (SOTA) works or matches their performance. The source code is publicly available at https://github.com/xmindflow/MSA-2Net., Comment: Accepted at BMVC 2024. Supplementary materials included at the end of the main paper (3 pages, 2 figures, 1 table)

We propose SlowFast-LLaVA (or SF-LLaVA for short), a training-free video large language model (LLM) that can jointly capture detailed spatial semantics and long-range temporal context without exceeding the token budget of commonly used LLMs. This is realized by using a two-stream SlowFast design of inputs for Video LLMs to aggregate features from sampled frames in an effective way. Specifically, the Slow pathway extracts features at a low frame rate while keeping as much spatial detail as possible (e.g., with 12x24 tokens), and the Fast pathway operates on a high frame rate but uses a larger spatial pooling stride (e.g., downsampling 6x) to focus on the motion cues. As a result, this design allows us to adequately capture both spatial and temporal features that are beneficial for detailed video understanding. Experimental results show that SF-LLaVA outperforms existing training-free methods on a wide range of video tasks. On some benchmarks, it achieves comparable or even better performance compared to state-of-the-art Video LLMs that are fine-tuned on video datasets. Code has been made available at: https://github.com/apple/ml-slowfast-llava., Comment: Technical report

Physics-inspired Generative Models (GMs), in particular Diffusion Models (DMs) and Poisson Flow Models (PFMs), enhance Bayesian methods and promise great utility in medical imaging. This review examines the transformative role of such generative methods. First, a variety of physics-inspired GMs, including Denoising Diffusion Probabilistic Models (DDPMs), Score-based Diffusion Models (SDMs), and Poisson Flow Generative Models (PFGMs and PFGM++), are revisited, with an emphasis on their accuracy, robustness as well as acceleration. Then, major applications of physics-inspired GMs in medical imaging are presented, comprising image reconstruction, image generation, and image analysis. Finally, future research directions are brainstormed, including unification of physics-inspired GMs, integration with Vision-Language Models (VLMs), and potential novel applications of GMs. Since the development of generative methods has been rapid, this review will hopefully give peers and learners a timely snapshot of this new family of physics-driven generative models and help capitalize their enormous potential for medical imaging.

The Ogden model is often considered as a standard model in the literature for application to the deformation of brain tissue. Here we show that, in some of those applications, the use of the Ogden model leads to non-convexity of the strain-energy function and mis-prediction of the correct concavity of the experimental stress-stretch curves over a range of the deformation domain. By contrast, we propose a family of models which provides a favourable fit to the considered datasets while remaining free from the highlighted shortcomings of the Ogden model. While, as we discuss, those shortcomings might be due to the artefacts of the testing protocols, the proposed family of models proves impervious to such artefacts.

A promising way to mitigate the expensive process of obtaining a high-dimensional signal is to acquire a limited number of low-dimensional measurements and solve an under-determined inverse problem by utilizing the structural prior about the signal. In this paper, we focus on adaptive acquisition schemes to save further the number of measurements. To this end, we propose a reinforcement learning-based approach that sequentially collects measurements to better recover the underlying signal by acquiring fewer measurements. Our approach applies to general inverse problems with continuous action spaces and jointly learns the recovery algorithm. Using insights obtained from theoretical analysis, we also provide a probabilistic design for our methods using variational formulation. We evaluate our approach on multiple datasets and with two measurement spaces (Gaussian, Radon). Our results confirm the benefits of adaptive strategies in low-acquisition horizon settings.

Workplace accidents continue to pose significant risks for human safety, particularly in industries such as construction and manufacturing, and the necessity for effective Personal Protective Equipment (PPE) compliance has become increasingly paramount. Our research focuses on the development of non-invasive techniques based on the Object Detection (OD) and Convolutional Neural Network (CNN) to detect and verify the proper use of various types of PPE such as helmets, safety glasses, masks, and protective clothing. This study proposes the SH17 Dataset, consisting of 8,099 annotated images containing 75,994 instances of 17 classes collected from diverse industrial environments, to train and validate the OD models. We have trained state-of-the-art OD models for benchmarking, and initial results demonstrate promising accuracy levels with You Only Look Once (YOLO)v9-e model variant exceeding 70.9% in PPE detection. The performance of the model validation on cross-domain datasets suggests that integrating these technologies can significantly improve safety management systems, providing a scalable and efficient solution for industries striving to meet human safety regulations and protect their workforce. The dataset is available at https://github.com/ahmadmughees/sh17dataset.

Preference alignment has become a crucial component in enhancing the performance of Large Language Models (LLMs), yet its impact in Multimodal Large Language Models (MLLMs) remains comparatively underexplored. Similar to language models, MLLMs for image understanding tasks encounter challenges like hallucination. In MLLMs, hallucination can occur not only by stating incorrect facts but also by producing responses that are inconsistent with the image content. A primary objective of alignment for MLLMs is to encourage these models to align responses more closely with image information. Recently, multiple works have introduced preference datasets for MLLMs and examined different alignment methods, including Direct Preference Optimization (DPO) and Proximal Policy Optimization (PPO). However, due to variations in datasets, base model types, and alignment methods, it remains unclear which specific elements contribute most significantly to the reported improvements in these works. In this paper, we independently analyze each aspect of preference alignment in MLLMs. We start by categorizing the alignment algorithms into two groups, offline (such as DPO), and online (such as online-DPO), and show that combining offline and online methods can improve the performance of the model in certain scenarios. We review a variety of published multimodal preference datasets and discuss how the details of their construction impact model performance. Based on these insights, we introduce a novel way of creating multimodal preference data called Bias-Driven Hallucination Sampling (BDHS) that needs neither additional annotation nor external models, and show that it can achieve competitive performance to previously published alignment work for multimodal models across a range of benchmarks.

Current multimodal and multitask foundation models like 4M or UnifiedIO show promising results, but in practice their out-of-the-box abilities to accept diverse inputs and perform diverse tasks are limited by the (usually rather small) number of modalities and tasks they are trained on. In this paper, we expand upon the capabilities of them by training a single model on tens of highly diverse modalities and by performing co-training on large-scale multimodal datasets and text corpora. This includes training on several semantic and geometric modalities, feature maps from recent state of the art models like DINOv2 and ImageBind, pseudo labels of specialist models like SAM and 4DHumans, and a range of new modalities that allow for novel ways to interact with the model and steer the generation, for example image metadata or color palettes. A crucial step in this process is performing discrete tokenization on various modalities, whether they are image-like, neural network feature maps, vectors, structured data like instance segmentation or human poses, or data that can be represented as text. Through this, we expand on the out-of-the-box capabilities of multimodal models and specifically show the possibility of training one model to solve at least 3x more tasks/modalities than existing ones and doing so without a loss in performance. This enables more fine-grained and controllable multimodal generation capabilities and allows us to study the distillation of models trained on diverse data and objectives into a unified model. We successfully scale the training to a three billion parameter model using tens of modalities and different datasets. The resulting models and training code are open sourced at 4m.epfl.ch., Comment: Project page at 4m.epfl.ch

Sequence modeling plays a vital role across various domains, with recurrent neural networks being historically the predominant method of performing these tasks. However, the emergence of transformers has altered this paradigm due to their superior performance. Built upon these advances, transformers have conjoined CNNs as two leading foundational models for learning visual representations. However, transformers are hindered by the $\mathcal{O}(N^2)$ complexity of their attention mechanisms, while CNNs lack global receptive fields and dynamic weight allocation. State Space Models (SSMs), specifically the \textit{\textbf{Mamba}} model with selection mechanisms and hardware-aware architecture, have garnered immense interest lately in sequential modeling and visual representation learning, challenging the dominance of transformers by providing infinite context lengths and offering substantial efficiency maintaining linear complexity in the input sequence. Capitalizing on the advances in computer vision, medical imaging has heralded a new epoch with Mamba models. Intending to help researchers navigate the surge, this survey seeks to offer an encyclopedic review of Mamba models in medical imaging. Specifically, we start with a comprehensive theoretical review forming the basis of SSMs, including Mamba architecture and its alternatives for sequence modeling paradigms in this context. Next, we offer a structured classification of Mamba models in the medical field and introduce a diverse categorization scheme based on their application, imaging modalities, and targeted organs. Finally, we summarize key challenges, discuss different future research directions of the SSMs in the medical domain, and propose several directions to fulfill the demands of this field. In addition, we have compiled the studies discussed in this paper along with their open-source implementations on our GitHub repository., Comment: This is the first version of our survey, and the paper is currently under review

In metals, low-energy effective theories are characterized by a set of coupling functions. Among them, the angle-dependent Fermi momentum specifies the size and shape of Fermi surface. Since the Fermi momentum grows incessantly under the renormalization group flow, a metallic fixed point is defined only modulo a rescaling of Fermi momentum. In this paper, we discuss the physical consequences of this projective nature of fixed points for non-Fermi liquids with hot Fermi surfaces. The first is the absence of a unique dynamical critical exponent that dictates the relative scaling between energy and momentum. The second is mismatches between the scaling dimensions of couplings and their relevancy. Nonetheless, each projective fixed point is characterized by a few marginal and relevant coupling functions, and the notion of universality survives. We illustrate our findings by charting the space of projective fixed points and extracting their universal properties for the Ising-nematic quantum critical metal beyond the patch theory. To control the theory, we use the dimensional regularization scheme that tunes the co-dimension of Fermi surface. Near the upper critical dimension, two exactly marginal coupling functions span the space of stable projective fixed points: functions that specify the shape of the Fermi surface and the angle-dependent Fermi velocity. All other coupling functions, including the Landau functions and the universal pairing interaction, are fixed by those two marginal functions. With decreasing dimensions, the forward scattering remains irrelevant while the pairing interaction becomes relevant near two dimensions. In two dimensions, it is expected that the universal superconducting fluctuations lower the symmetry of the non-Fermi liquid realized above the superconducting transition temperatures from the loop U(1) group to a proper subgroup., Comment: 71 pages; v3) Fig. 10b replaced

The rise of Transformer architectures has revolutionized medical image segmentation, leading to hybrid models that combine Convolutional Neural Networks (CNNs) and Transformers for enhanced accuracy. However, these models often suffer from increased complexity and overlook the interplay between spatial and channel features, which is vital for segmentation precision. We introduce LHU-Net, a streamlined Hybrid U-Net for volumetric medical image segmentation, designed to first analyze spatial and then channel features for effective feature extraction. Tested on five benchmark datasets (Synapse, LA, Pancreas, ACDC, BRaTS 2018), LHU-Net demonstrated superior efficiency and accuracy, notably achieving a 92.66 Dice score on ACDC with 85\% fewer parameters and a quarter of the computational demand compared to leading models. This performance, achieved without pre-training, extra data, or model ensembles, sets new benchmarks for computational efficiency and accuracy in segmentation, using under 11 million parameters. This achievement highlights that balancing computational efficiency with high accuracy in medical image segmentation is feasible. Our implementation of LHU-Net is freely accessible to the research community on GitHub (https://github.com/xmindflow/LHUNet).

Intrigued by the inherent ability of the human visual system to identify salient regions in complex scenes, attention mechanisms have been seamlessly integrated into various Computer Vision (CV) tasks. Building upon this paradigm, Vision Transformer (ViT) networks exploit attention mechanisms for improved efficiency. This review navigates the landscape of redesigned attention mechanisms within ViTs, aiming to enhance their performance. This paper provides a comprehensive exploration of techniques and insights for designing attention mechanisms, systematically reviewing recent literature in the field of CV. This survey begins with an introduction to the theoretical foundations and fundamental concepts underlying attention mechanisms. We then present a systematic taxonomy of various attention mechanisms within ViTs, employing redesigned approaches. A multi-perspective categorization is proposed based on their application, objectives, and the type of attention applied. The analysis includes an exploration of the novelty, strengths, weaknesses, and an in-depth evaluation of the different proposed strategies. This culminates in the development of taxonomies that highlight key properties and contributions. Finally, we gather the reviewed studies along with their available open-source implementations at our \href{https://github.com/mindflow-institue/Awesome-Attention-Mechanism-in-Medical-Imaging}{GitHub}\footnote{\url{https://github.com/xmindflow/Awesome-Attention-Mechanism-in-Medical-Imaging}}. We aim to regularly update it with the most recent relevant papers., Comment: Submitted to Computational Visual Media Journal

Soft materials such as rubbers, silicones, gels and biological tissues have a nonlinear response to large deformations, a phenomenon which in principle can be captured by hyperelastic models. The suitability of a candidate hyperelastic strain energy function is then determined by comparing its predicted response to the data gleaned from tests and adjusting the material parameters to get a good fit, an exercise which can be deceptive because of nonlinearity. Here we propose to generalise the approach of Rivlin and Saunders [Phil Trans A 243 (1951) 251-288] who, instead of reporting the data as stress against stretch, manipulated these measures to create the 'Mooney plot', where the Mooney-Rivlin model is expected to produce a linear fit. We show that extending this idea to other models and modes of deformation (tension, shear, torsion, etc.) is advantageous, not only (a) for the fitting procedure, but also to (b) delineate trends in the deformation which are not obvious from the raw data (and may be interpreted in terms of micro-, meso-, and macro-structures) and (c) obtain a bounded condition number \k{appa} over the whole range of deformation; a robustness which is lacking in other plots and spaces.

We investigate properties of the transmission amplitude of quantum graphs and microwave networks composed of regular polygons such as triangles and squares. We show that for the graphs composed of regular polygons with the edges of the length $l$ the transmission amplitude displays a band of transmission suppression with some narrow peaks of full transmission. The peaks are distributed symmetrically with respect to the symmetry axis $kl=\pi$, where $k$ is the wave vector. For microwave networks the transmission peak amplitudes are reduced and their symmetry is broken due to the influence of internal absorption. We demonstrate that for the graphs composed of the same polygons but separated by the edges of length $l' < l$ the transmission spectrum is generally not symmetric according to the axis $kl'=\pi$. We also show that graphs composed of regular polygons of different size with the edges being irrational numbers are not fully chaotic and their level spacing distribution and the spectral rigidity are well described by the Berry-Robnik distributions. Moreover, the transmission spectrum of such a graph displays peaks which are very close to $1$. Furthermore, the microwave networks are investigated in the time-domain using short Gaussian pulses. In this case the delay-time distributions, though very sensitive to the internal structure of the networks, show the sequences of transmitted peaks with the amplitudes much smaller than the input one. The analyzed properties of the graphs and networks suggest that they can be effectively used to manipulate quantum and wave transport.

A Dielectric Barrier Discharge (DBD) plasma actuator for controlling airflow is proposed. It consists of diverging and converging nozzles, two concentric cylinders and an actuator mounted in-between the two cylinders. The actuator employs electrohydrodynamic (EHD) body force to induce an air jet within the air gap between the two cylinders, effectively creating a suction area while passing through the diverging nozzle, due to the Coanda effect. While merging with the air stream inside the inner cylinder, the Coanda jet effectively enhances amplification of the airflow. The outflow rate is measured by a velocity sensor at the outlet and controlled by the plasma actuator. The control strategy is based on the Active Disturbance Rejection Control (ADRC) and compared to the baseline PID controller. The actuator was modelled by seamlessly linking two modeling platforms for a co-simulation study. The CFD simulation of the plasma and airflow was carried out in the COMSOL multi-physics commercial software, and the control was implemented in the Simulink. The DBD plasma model was based on the two-species model of discharge, and the electric body force, calculated from the plasma simulation, was used in the Navier-Stokes equation for the turbulent flow simulation. The plasma-air flow system was analyzed using the input (the actuator voltage) and output (the outlet flow rate) data for the control design. Finally, the performance of the system of air flow control device was tested and discussed in the co-simulation process.