Papers

Millimeter-wave (mmWave) communication offers vast bandwidth for next-generation wireless systems but faces severe path loss, Doppler effects, and hardware impairments. Orthogonal Time Frequency Space (OTFS) modulation has emerged as a robust waveform for high-mobility and doubly dispersive channels, outperforming OFDM under strong Doppler. However, the most studied multicarrier OTFS (MC-OTFS) is not easily predictable because the input-output (I$/$O) relation is not given by (twisted) convolution. Recently, the Zak-transform based OTFS (Zak-OTFS or OTFS 2$.$0) was proposed, which provides a single domain delay Doppler (DD) processing framework with predictable I$/$O behavior. This paper presents one of the first over-the-air (OTA) demonstrations of Zak-OTFS at mmWave frequencies. We design a complete Zak-OTFS based mmWave OTA system featuring root-raised-cosine (RRC) filtering for enhanced DD-domain predictability, higher-order modulations up to 16-QAM, and a low-overhead preamble for synchronization. A comprehensive signal model incorporating carrier frequency offset (CFO) and timing impairments is developed, showing these effects can be jointly captured within the effective DD-domain channel. Experimental validation on the COSMOS testbed confirms the feasibility and robustness of Zak-OTFS under realistic mmWave conditions, highlighting its potential for efficient implementations in beyond-5G and 6G systems.

This thesis presents the development of ARGUS, a framework for mission planning for Unmanned Ground Vehicles (UGVs) in tactical environments. The system is designed to translate battlefield complexity and the commander's intent into executable action plans. To this end, ARGUS employs a processing pipeline that takes as input geospatial terrain data, military intelligence on existing threats and their probable locations, and mission priorities defined by the commander. Through a set of integrated modules, the framework processes this information to generate optimized trajectories that balance mission objectives against the risks posed by threats and terrain characteristics. A fundamental capability of ARGUS is its dynamic nature, which allows it to adapt plans in real-time in response to unforeseen events, reflecting the fluid nature of the modern battlefield. The system's interoperability were validated in a practical exercise with the Portuguese Army, where it was successfully demonstrated that the routes generated by the model can be integrated and utilized by UGV control systems. The result is a decision support tool that not only produces an optimal trajectory but also provides the necessary insights for its execution, thereby contributing to greater effectiveness and safety in the employment of autonomous ground systems.

In computational pathology, the gigapixel scale of Whole-Slide Images (WSIs) necessitates their division into thousands of smaller patches. Analyzing these high-dimensional patch embeddings is computationally expensive and risks diluting key diagnostic signals with many uninformative patches. Existing patch selection methods often rely on random sampling or simple clustering heuristics and typically fail to explicitly manage the crucial trade-off between the number of selected patches and the accuracy of the resulting slide representation. To address this gap, we propose EvoPS (Evolutionary Patch Selection), a novel framework that formulates patch selection as a multi-objective optimization problem and leverages an evolutionary search to simultaneously minimize the number of selected patch embeddings and maximize the performance of a downstream similarity search task, generating a Pareto front of optimal trade-off solutions. We validated our framework across four major cancer cohorts from The Cancer Genome Atlas (TCGA) using five pretrained deep learning models to generate patch embeddings, including both supervised CNNs and large self-supervised foundation models. The results demonstrate that EvoPS can reduce the required number of training patch embeddings by over 90% while consistently maintaining or even improving the final classification F1-score compared to a baseline that uses all available patches' embeddings selected through a standard extraction pipeline. The EvoPS framework provides a robust and principled method for creating efficient, accurate, and interpretable WSI representations, empowering users to select an optimal balance between computational cost and diagnostic performance.

This paper aims to proactively diagnose and manage frequency instability risks from a steady-state perspective, without the need for derivative-dependent transient modeling. Specifically, we jointly address two questions (Q1) Survivability: following a disturbance and the subsequent primary frequency response, can the system settle into a healthy steady state (feasible with an acceptable frequency deviation $\Delta f$)? (Q2) Dominant Vulnerability: if found unstable, what critical vulnerabilities create instability and/or full collapse? To address these questions, we first augment steady-state power flow states to include frequency-dependent governor relationships (i.e., governor power flow). Afterwards, we propose a frequency-aware sparse optimization that finds the minimal set of bus locations with measurable compensations (corrective actions) to enforce power balance and maintain frequency within predefined/acceptable bounds. We evaluate our method on standard transmission systems to empirically validate its ability to localize dominant sources of vulnerabilities. For a 1354-bus large system, our method detects compensations to only four buses under N-1 generation outage (3424.8 MW) while enforcing a maximum allowable steady-state frequency drop of 0.06 Hz (otherwise, frequency drops by nearly 0.08 Hz). We further validate the scalability of our method, requiring less than four minutes to obtain sparse solutions for the 1354-bus system.

We study the problem of learning policies that maximize cumulative reward while satisfying safety constraints, even when the real environment differs from a simulator or nominal model. We focus on robust constrained Markov decision processes (RCMDPs), where the agent must maximize reward while ensuring cumulative utility exceeds a threshold under the worst-case dynamics within an uncertainty set. While recent works have established finite-time iteration complexity guarantees for RCMDPs using policy optimization, their sample complexity guarantees remain largely unexplored. In this paper, we first show that Markovian policies may fail to be optimal even under rectangular uncertainty sets unlike the unconstrained robust MDP. To address this, we introduce an augmented state space that incorporates the remaining utility budget into the state representation. Building on this formulation, we propose a novel Robust constrained Value iteration (RCVI) algorithm with a sample complexity of $\mathcal{\tilde{O}}(|S||A|H^5/\epsilon^2)$ achieving at most $\epsilon$ violation using a generative model where $|S|$ and $|A|$ denote the sizes of the state and action spaces, respectively, and $H$ is the episode length. To the best of our knowledge, this is the first sample complexity guarantee for RCMDP. Empirical results further validate the effectiveness of our approach.

There have been significant advances in recent years in the development of forward error correction decoders that can decode codes of any structure, including practical realizations in synthesized circuits and taped out chips. While essentially all soft-decision decoders assume that bits have been impacted independently on the channel, for one of these new approaches it has been established that channel dependencies can be exploited to achieve superior decoding accuracy, resulting in Ordered Reliability Bits Guessing Random Additive Noise Decoding Approximate Independence (ORBGRAND-AI). Building on that capability, here we consider the integration of ORBGRAND-AI as a pattern generator for Guessing Codeword Decoding (GCD). We first establish that a direct approach delivers mildly degraded block error rate (BLER) but with reduced number of queried patterns when compared to ORBGRAND-AI. We then show that with a more nuanced approach it is possible to leverage total correlation to deliver an additional BLER improvement of ~0.75$ dB while retaining reduced query numbers.

We study a two-player dynamic Stackelberg game where the follower's intention is unknown to the leader. Classical formulations of the Stackelberg equilibrium (SE) assume that the follower's best response (BR) function is known to the leader. However, this is not always true in practice. We study a setting in which the leader receives updated beliefs about the follower BR before the end of the game, such that the update prompts the leader and subsequently the follower to re-optimize their strategies. We characterize the optimality guarantees of the SE solutions under this belief update for both open loop and feedback information structures. Interestingly, we prove that in general, assuming an incorrect follower's BR may lead to a lower leader cost over the entire game than knowing the true follower's BR. We support these results with numerical examples in a linear quadratic (LQ) Stackelberg game, and use Monte Carlo simulations to show that the instances of incorrect BR achieving lower leader costs are non-trivial in collision avoidance LQ Stackelberg games.

This paper presents a novel modeling approach for unsteady aircraft airflow, leveraging the Lorenz attractor framework. The proposed model is based on the force distribution exerted by a lift-generating wing on the surrounding fluid. It distinguishes between turbulent and nominal components of the force distribution, with the nominal force distribution modeled to peak at the wing and decay linearly into the free stream. This separation allows the turbulent component to be represented by a transport equation that is influenced by flight conditions, specifically dynamic pressure and angle of attack. Consequently, the Navier-Stokes equations, along with the turbulence transport equation, can be transformed into a reduced-order model characterized by three scalar ordinary differential equations - similar to the Lorenz attractor. This resulting system effectively captures chaotic behavior, facilitating the exploration of complex dynamics without the computational demands of solving the full Navier-Stokes equations. A simulation trade study is conducted that models wing rock phenomena at high angles of attack, demonstrating the effectiveness of the proposed approach in capturing the intricate dynamics of unsteady aircraft aerodynamics.

This paper presents a systematic approach for designing robust linear proportional-integral (PI) servo-controllers that effectively manage control input and output constraints in flight control systems. The control design leverages the Nagumo Theorem and the Comparison Lemma to prove constraint satisfaction, while employing min-norm optimal controllers in a manner akin to Control Barrier Functions. This results in a continuous piecewise-linear state feedback policy that maintains the analyzability of the closed-loop system through the principles of linear systems theory. Additionally, we derive multi-input multi-output (MIMO) robustness margins, demonstrating that our approach enables robust tracking of external commands even in the presence of operational constraints. Moreover, the proposed control design offers a systematic approach for anti-windup protection. Through flight control trade studies, we illustrate the applicability of the proposed framework to real-world safety-critical aircraft control scenarios. Notably, MIMO margin analysis with active constraints reveals that our method preserves gain and phase margins comparable to those of the unconstrained case, in contrast to controllers that rely on hard saturation heuristics, which suffer significant performance degradation under active constraints. Simulation results using a nonlinear six-degree-of-freedom rigid body aircraft model further validate the effectiveness of our method in achieving constraint satisfaction, robustness, and effective anti-windup protection.

This paper introduces Roundabout Constrained Convex Generators (RCGs), a set representation framework for modeling multiply connected regions in control and verification applications. The RCG representation extends the constrained convex generators framework by incorporating an inner exclusion zone, creating sets with topological holes that naturally arise in collision avoidance and safety-critical control problems. We present two equivalent formulations: a set difference representation that provides geometric intuition and a unified parametric representation that facilitates computational implementation. The paper establishes closure properties under fundamental operations, including linear transformations, Minkowski sums, and intersections with convex generator sets. We derive special cases, including roundabout zonotopes and roundabout ellipsotopes, which offer computational advantages for specific norm selections. The framework maintains compatibility with existing optimization solvers while enabling the representation of non-convex feasible regions that were previously challenging to model efficiently.

The feasibility and cost-effectiveness of continued growth in solar photovoltaics are closely tied to siting decisions. But trade-offs between costs and technical potential between land categories, especially brownfields and rooftop sites, have not been quantified, despite increasing resistance to and policy interest in reducing use of greenfield sites (e.g., prime agricultural lands). We examine the effect of siting decisions across land types for utility-scale and rooftop PV on the feasibility and cost of meeting solar deployment targets across the Eastern U.S. We build a database of solar PV supply curves by land type for each county in the Eastern Interconnect (EI) region (~2,400 counties). Our supply curves quantify technical potential versus levelized cost across greenfield, brownfield, and rooftop land types. With these supply curves and a 2035 solar deployment target (435 GW) aligned with a decarbonized power system, we quantify cost and capacity trade-offs using scenarios that prioritize solar PV deployment on different land types. We find greenfield, particularly prime agriculture, sites offer the lowest levelized costs for meeting capacity targets, of 39 to 57 $/MWh. Contaminated lands, often prioritized in policy to reduce land use conflict, have limited technical potential and impose a cost premium of 14-33% relative to greenfield sites. Rooftop PV provides enough technical potential for meeting capacity targets but comes at consistently higher costs, with minimum LCOEs of roughly 70 $/MWh or well above the highest-cost greenfield sites. Our results detail heterogeneous siting trade-offs across the Eastern United States, enabling targeted policy design to meet deployment targets while balancing costs and land use conflicts.

The growing demand for efficient delivery of common content to multiple user equipments (UEs) has motivated significant research in physical-layer multicasting. By exploiting the beamforming capabilities of massive MIMO, multicasting provides a spectrum-efficient solution that avoids unnecessary intra-group interference. A key challenge, however, is solving the max-min fair (MMF) and quality-of-service (QoS) multicast beamforming optimization problems, which are NP-hard due to the non-convex structure and the requirement for rank-1 solutions. Traditional approaches based on semidefinite relaxation (SDR) followed by randomization exhibit poor scalability with system size, while state-of-the-art successive convex approximation (SCA) methods only guarantee convergence to stationary points. In this paper, we propose an alternating direction method of multipliers (ADMM)-based framework for MMF and QoS multicast beamforming in cell-free massive MIMO networks. The algorithm leverages SDR but incorporates a novel iterative elimination strategy within the ADMM updates to efficiently obtain near-global optimal rank-1 beamforming solutions with reduced computational complexity compared to standard SDP solvers and randomization methods. Numerical evaluations demonstrate that the proposed ADMM-based procedure not only achieves superior spectral efficiency but also scales favorably with the number of antennas and UEs compared to state-of-the-art SCA-based algorithms, making it a practical tool for next-generation multicast systems.

The prevalence of user-generated content (UGC) on platforms such as YouTube and TikTok has rendered no-reference (NR) perceptual video quality assessment (VQA) vital for optimizing video delivery. Nonetheless, the characteristics of non-professional acquisition and the subsequent transcoding of UGC video on sharing platforms present significant challenges for NR-VQA. Although NR-VQA models attempt to infer mean opinion scores (MOS), their modeling of subjective scores for compressed content remains limited due to the absence of fine-grained perceptual annotations of artifact types. To address these challenges, we propose CAMP-VQA, a novel NR-VQA framework that exploits the semantic understanding capabilities of large vision-language models. Our approach introduces a quality-aware prompting mechanism that integrates video metadata (e.g., resolution, frame rate, bitrate) with key fragments extracted from inter-frame variations to guide the BLIP-2 pretraining approach in generating fine-grained quality captions. A unified architecture has been designed to model perceptual quality across three dimensions: semantic alignment, temporal characteristics, and spatial characteristics. These multimodal features are extracted and fused, then regressed to video quality scores. Extensive experiments on a wide variety of UGC datasets demonstrate that our model consistently outperforms existing NR-VQA methods, achieving improved accuracy without the need for costly manual fine-grained annotations. Our method achieves the best performance in terms of average rank and linear correlation (SRCC: 0.928, PLCC: 0.938) compared to state-of-the-art methods. The source code and trained models, along with a user-friendly demo, are available at: https://github.com/xinyiW915/CAMP-VQA.

Although a variety of transformers have been proposed for symbolic music generation in recent years, there is still little comprehensive study on how specific design choices affect the quality of the generated music. In this work, we systematically compare different datasets, model architectures, model sizes, and training strategies for the task of symbolic piano music generation. To support model development and evaluation, we examine a range of quantitative metrics and analyze how well they correlate with human judgment collected through listening studies. Our best-performing model, a 950M-parameter transformer trained on 80K MIDI files from diverse genres, produces outputs that are often rated as human-composed in a Turing-style listening survey.

Large language models (LLMs) have recently achieved impressive results in speech recognition across multiple modalities, including Auditory Speech Recognition (ASR), Visual Speech Recognition (VSR), and Audio-Visual Speech Recognition (AVSR). Despite this progress, current LLM-based approaches typically address each task independently, training separate models that raise computational and deployment resource use while missing potential cross-task synergies. They also rely on fixed-rate token compression, which restricts flexibility in balancing accuracy with efficiency. These limitations highlight the need for a unified framework that can support ASR, VSR, and AVSR while enabling elastic inference. To this end, we present Omni-AVSR, a unified audio-visual LLM that combines efficient multi-granularity training with parameter-efficient adaptation. Specifically, we adapt the matryoshka representation learning paradigm to efficiently train across multiple audio and visual granularities, reducing its inherent training resource use. Furthermore, we explore three LoRA-based strategies for adapting the backbone LLM, balancing shared and task-specific specialization. Experiments on LRS2 and LRS3 show that Omni-AVSR achieves comparable or superior accuracy to state-of-the-art baselines while training a single model at substantially lower training and deployment resource use. The model also remains robust under acoustic noise, and we analyze its scaling behavior as LLM size increases, providing insights into the trade-off between performance and efficiency.

Mobility-on-demand systems like ride-hailing have transformed urban transportation, but they have also exacerbated socio-economic inequalities in access to these services, also due to surge pricing strategies. Although several fairness-aware frameworks have been proposed in smart mobility, they often overlook the temporal and situational variability of user urgency that shapes real-world transportation demands. This paper introduces a non-monetary, Karma-based mechanism that models endogenous urgency, allowing user time-sensitivity to evolve in response to system conditions as well as external factors. We develop a theoretical framework maintaining the efficiency and fairness guarantees of classical Karma economies, while accommodating this realistic user behavior modeling. Applied to a simplified simulated mobility-on-demand scenario, we provide a proof-of-concept illustration of the proposed framework, showing that it exhibits promising behavior in terms of system efficiency and equitable resource allocation, while acknowledging that a full treatment of realistic MoD complexity remains an important direction for future work.

Beamforming with desired directivity patterns using compact microphone arrays is essential in many audio applications. Directivity patterns achievable using traditional beamformers depend on the number of microphones and the array aperture. Generally, their effectiveness degrades for compact arrays. To overcome these limitations, we propose a neural directional filtering (NDF) approach that leverages deep neural networks to enable sound capture with a predefined directivity pattern. The NDF computes a single-channel complex mask from the microphone array signals, which is then applied to a reference microphone to produce an output that approximates a virtual directional microphone with the desired directivity pattern. We introduce training strategies and propose data-dependent metrics to evaluate the directivity pattern and directivity factor. We show that the proposed method: i) achieves a frequency-invariant directivity pattern even above the spatial aliasing frequency, ii) can approximate diverse and higher-order patterns, iii) can steer the pattern in different directions, and iv) generalizes to unseen conditions. Lastly, experimental comparisons demonstrate superior performance over conventional beamforming and parametric approaches.

Low-altitude economy (LAE) is rapidly emerging as a key driver of innovation, encompassing economic activities taking place in airspace below 500 meters. Unmanned aerial vehicles (UAVs) provide valuable tools for logistics collection within LAE systems, offering the ability to navigate through complex environments, avoid obstacles, and improve operational efficiency. However, logistics collection tasks involve UAVs flying through complex three-dimensional (3D) environments while avoiding obstacles, where traditional UAV trajectory design methods,typically developed under free-space conditions without explicitly accounting for obstacles, are not applicable. This paper presents, we propose a novel algorithm that combines the Lin-Kernighan-Helsgaun (LKH) and Deep Deterministic Policy Gradient (DDPG) methods to minimize the total collection time. Specifically, the LKH algorithm determines the optimal order of item collection, while the DDPG algorithm designs the flight trajectory between collection points. Simulations demonstrate that the proposed LKH-DDPG algorithm significantly reduces collection time by approximately 49 percent compared to baseline approaches, thereby highlighting its effectiveness in optimizing UAV trajectories and enhancing operational efficiency for logistics collection tasks in the LAE paradigm.

Emerging 6G wireless systems suffer severe performance degradation in challenging environments like high-speed trains traversing dense urban corridors and Unmanned Aerial Vehicles (UAVs) links over mountainous terrain. These scenarios exhibit non-Gaussian, non-stationary channels with heavy-tailed fading and abrupt signal fluctuations. To address these challenges, this paper proposes a novel wireless channel model based on symmetric $\alpha$-stable Lévy processes, thereby enabling continuous-time state-space characterization of both long-term and short-term fading. Building on this model, a generalized optimal control framework is developed via a fractional Hamilton-Jacobi-Bellman (HJB) equation that incorporates the Riesz fractional operator to capture non-local spatial effects and memory-dependent dynamics. The existence and uniqueness of viscosity solutions to the fractional HJB equation are rigorously established, thus ensuring the theoretical validity of the proposed control formulation. Numerical simulations conducted in a multi-cell, multi-user downlink setting demonstrate the effectiveness of the fractional HJB-based strategy in optimizing transmission power under heavy-tailed co-channel and multi-user interference.

The rapid growth of data centres poses an evolving challenge for power systems with high variable renewable energy. Traditionally operated as passive electrical loads, data centres, have the potential to become active participants that provide flexibility to the grid. However, quantifying and utilising this flexibility have not yet been fully explored. This paper presents an integrated, whole facility optimisation model to investigate the least cost operating schedule of data centres and characterise the aggregate flexibility available from data centres to the power system. The model accounts for IT workload shifting, UPS energy storage, and cooling system. Motivated by the need to alleviate the increasing strain on power systems while leveraging their untapped flexibility potential, this study makes two primary contributions: (i) an operational optimisation model that integrates IT scheduling, UPS operation, and cooling dynamics to establish a cost optimal baseline operation, and (ii) a duration-aware flexibility assessment that, for any given start time and power deviation, computes the maximum feasible duration from this baseline while respecting all operational, thermal, and recovery constraints. This method characterises the aggregate flexibility envelope. Results reveal a clear temporal structure and a notable asymmetry in flexibility provision: upward flexibility (electricity load reduction) is driven by deferring IT workload, which allows for a secondary reduction in cooling power. Downward flexibility (electricity load increase) relies on increasing power consumption of the cooling system, supported by the TES buffer, and charging the UPS. This framework translates abstract flexibility potential into quantified flexibility magnitude and duration that system operators could investigate for use in services such as reserve, frequency response, and price responsive demand.

Recent advances in latent diffusion models have demonstrated state-of-the-art performance in high-dimensional time-series data synthesis while providing flexible control through conditioning and guidance. However, existing methodologies primarily rely on musical context or natural language as the main modality of interacting with the generative process, which may not be ideal for expert users who seek precise fader-like control over specific musical attributes. In this work, we explore the application of denoising diffusion processes as plug-and-play latent constraints for unconditional symbolic music generation models. We focus on a framework that leverages a library of small conditional diffusion models operating as implicit probabilistic priors on the latents of a frozen unconditional backbone. While previous studies have explored domain-specific use cases, this work, to the best of our knowledge, is the first to demonstrate the versatility of such an approach across a diverse array of musical attributes, such as note density, pitch range, contour, and rhythm complexity. Our experiments show that diffusion-driven constraints outperform traditional attribute regularization and other latent constraints architectures, achieving significantly stronger correlations between target and generated attributes while maintaining high perceptual quality and diversity.

Voice conversion models modify timbre while preserving paralinguistic features, enabling applications like dubbing and identity protection. However, most VC systems require access to target utterances, limiting their use when target data is unavailable or when users desire conversion to entirely novel, unseen voices. To address this, we introduce a lightweight method SpeakerVAE to generate novel speakers for VC. Our approach uses a deep hierarchical variational autoencoder to model the speaker timbre space. By sampling from the trained model, we generate novel speaker representations for voice synthesis in a VC pipeline. The proposed method is a flexible plug-in module compatible with various VC models, without co-training or fine-tuning of the base VC system. We evaluated our approach with state-of-the-art VC models: FACodec and CosyVoice2. The results demonstrate that our method successfully generates novel, unseen speakers with quality comparable to that of the training speakers.

Explicit latent variable models provide a flexible yet powerful framework for data synthesis, enabling controlled manipulation of generative factors. With latent variables drawn from a tractable probability density function that can be further constrained, these models enable continuous and semantically rich exploration of the output space by navigating their latent spaces. Structured latent representations are typically obtained through the joint minimization of regularization loss functions. In variational information bottleneck models, reconstruction loss and Kullback-Leibler Divergence (KLD) are often linearly combined with an auxiliary Attribute-Regularization (AR) loss. However, balancing KLD and AR turns out to be a very delicate matter. When KLD dominates over AR, generative models tend to lack controllability; when AR dominates over KLD, the stochastic encoder is encouraged to violate the standard normal prior. We explore this trade-off in the context of symbolic music generation with explicit control over continuous musical attributes. We show that existing approaches struggle to jointly minimize both regularization objectives, whereas suitable attribute transformations can help achieve both controllability and regularization of the target latent dimensions.

Nonnegative matrix factorization (NMF) is a linear dimensionality reduction technique for nonnegative data, with applications such as hyperspectral unmixing and topic modeling. NMF is a difficult problem in general (NP-hard), and its solutions are typically not unique. To address these two issues, additional constraints or assumptions are often used. In particular, separability assumes that the basis vectors in the NMF are equal to some columns of the input matrix. In that case, the problem is referred to as separable NMF (SNMF) and can be solved in polynomial-time with robustness guarantees, while identifying a unique solution. However, in real-world scenarios, due to noise or variability, multiple data points may lie near the basis vectors, which SNMF does not leverage. In this work, we rely on the smooth separability assumption, which assumes that each basis vector is close to multiple data points. We explore the properties of the corresponding problem, referred to as smooth SNMF (SSNMF), and examine how it relates to SNMF and orthogonal NMF. We then propose a convex model for SSNMF and show that it provably recovers the sought-after factors, even in the presence of noise. We finally adapt an existing fast gradient method to solve this convex model for SSNMF, and show that it compares favorably with state-of-the-art methods on both synthetic and hyperspectral datasets.

Deep learning-based low-dose computed tomography reconstruction methods already achieve high performance on standard image quality metrics like peak signal-to-noise ratio and structural similarity index measure. Yet, they frequently fail to preserve the critical anatomical details needed for diagnostic tasks. This fundamental limitation hinders their clinical applicability despite their high metric scores. We propose a novel task-adaptive reconstruction framework that addresses this gap by incorporating a frozen pre-trained task network as a regularization term in the reconstruction loss function. Unlike existing joint-training approaches that simultaneously optimize both reconstruction and task networks, and risk diverging from satisfactory reconstructions, our method leverages a pre-trained task model to guide reconstruction training while still maintaining diagnostic quality. We validate our framework on a liver and liver tumor segmentation task. Our task-adaptive models achieve Dice scores up to 0.707, approaching the performance of full-dose scans (0.874), and substantially outperforming joint-training approaches (0.331) and traditional reconstruction methods (0.626). Critically, our framework can be integrated into any existing deep learning-based reconstruction model through simple loss function modification, enabling widespread adoption for task-adaptive optimization in clinical practice. Our codes are available at: https://github.com/itu-biai/task_adaptive_ct

Background parenchymal enhancement (BPE) on breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) shows potential as a breast cancer risk marker. Clinically, BPE is qualitatively assessed by radiologists, but quantitative BPE measures offer potential for more precise risk evaluation. This study evaluated an existing open-source, fully-automated deep learning-based (DL-based) method for segmenting fibroglandular tissue (FGT) to quantify BPE and compared it to a semi-automated fuzzy c-means method. Using breast MRI examinations from 100 women, we evaluated segmentation agreement, concordance across quantitative BPE metrics, and associations with qualitative BPE. The quality of FGT segmentations from both methods was scored by a radiologist. While the DL-based and semi-automated methods showed good agreement for quantitative BPE measurements, DL-based measures more strongly correlated with qualitative BPE assessments and DL-based segmentations were scored as higher quality by the radiologist. Our findings suggest that DL-based FGT segmentation enhances efficiency for objective BPE quantification and may improve standardized breast cancer risk assessment.

Deploying lightweight medical image segmentation models on edge devices presents two major challenges: 1) efficiently handling the stark contrast between lesion boundaries and background regions, and 2) the sharp drop in accuracy that occurs when pursuing extremely lightweight designs (e.g., <0.5M parameters). To address these problems, this paper proposes TauFlow, a novel lightweight segmentation model. The core of TauFlow is a dynamic feature response strategy inspired by brain-like mechanisms. This is achieved through two key innovations: the Convolutional Long-Time Constant Cell (ConvLTC), which dynamically regulates the feature update rate to "slowly" process low-frequency backgrounds and "quickly" respond to high-frequency boundaries; and the STDP Self-Organizing Module, which significantly mitigates feature conflicts between the encoder and decoder, reducing the conflict rate from approximately 35%-40% to 8%-10%.

The objective of this paper is to investigate graph-theoretic conditions for structural herdability of an LTI system. In particular, we are interested in the structural sign (SS) herdability of a system wherein the underlying digraph representing it is signed. Structural herdability finds applications in various domains like power networks, biological networks, opinion dynamics, multi-robot shepherding, etc. We begin the analysis by introducing a layered graph representation Gs of the signed digraph G; such a representation allows us to capture the signed distances between the nodes with ease. We construct a subgraph of G_s that characterizes paths of identical signs between layers and uniform path lengths, referred to as a layer-wise unisigned graph LUG(G_s). A special subgraph of an LUG(G_s), denoted as an LUG^H(G_s), is key to achieving SS herdability. This is because we prove that an LTI system is SS herdable if and only if there exists an LUG^H(G_s) which covers all the nodes of the given digraph. To the best of our knowledge, such a graphical test is one of the first methods which allows us to check SS herdability for arbitrary digraph topologies. Interestingly, the analysis also reveals that a system can be SS herdable even in the presence of (signed and layer) dilation in the associated digraph (note that such a behaviour has been shown to be impossible in directed trees). Additionally, we also extend these results to digraphs with multiple leader and driver nodes. In order to illustrate all the results, we present numerous examples throughout the paper.

The growing integration of renewable energy sources (RESs) in modern power systems has intensified the need for resilient and efficient microgrid solutions. DC microgrids have gained prominence due to their reduced conversion losses, simplified interfacing with DC-based RESs, and improved reliability. To manage the inherent variability of RESs and ensure stable operation, energy management systems (EMS) have become essential. While various EMS algorithms have been proposed and validated using real-time simulation platforms, most assume ideal communication conditions or rely on simplified network models, overlooking the impact of realistic communication delays on EMS performance. This paper presents a novel real-time cyber-physical system (CPS) testbed for evaluating EMS performance in DC microgrids under realistic communication delays. The proposed testbed integrates a DC microgrid modeled in OPAL-RT with an EMS controller implemented on a Raspberry Pi (RPi). The communication network is emulated using EXataCPS, enabling the exchange of actual power system traffic and the replication of realistic latency conditions. This comprehensive setup captures the interplay between power system dynamics, EMS control, and communication network behavior.

Early cancer detection is crucial for improving patient outcomes, and 18F FDG PET/CT imaging plays a vital role by combining metabolic and anatomical information. Accurate lesion detection remains challenging due to the need to identify multiple lesions of varying sizes. In this study, we investigate the effect of adding anatomy prior information to deep learning-based lesion detection models. In particular, we add organ segmentation masks from the TotalSegmentator tool as auxiliary inputs to provide anatomical context to nnDetection, which is the state-of-the-art for lesion detection, and Swin Transformer. The latter is trained in two stages that combine self-supervised pre-training and supervised fine-tuning. The method is tested in the AutoPET and Karolinska lymphoma datasets. The results indicate that the inclusion of anatomical priors substantially improves the detection performance within the nnDetection framework, while it has almost no impact on the performance of the vision transformer. Moreover, we observe that Swin Transformer does not offer clear advantages over conventional convolutional neural network (CNN) encoders used in nnDetection. These findings highlight the critical role of the anatomical context in cancer lesion detection, especially in CNN-based models.