Machine learning framework for CDL channel profile estimation and channel reconstruction from limited CSI feedback

Objective:Real-time cone-beam computed tomography (CBCT) provides instantaneous visualization of patient anatomy for image guidance, motion tracking, and online treatment adaptation in radiotherapy. While many real-time imaging and motion tracking methods leveraged patient-specific prior information to alleviate under-sampling challenges and meet the temporal constraint (< 500 ms), the prior information can be outdated and introduce biases, thus compromising the imaging and motion tracking accuracy. To address this challenge, we developed a framework (DREME) for real-time CBCT imaging and motion estimation, without relying on patient-specific prior knowledge.Approach:DREME incorporates a deep learning-based real-time CBCT imaging and motion estimation method into a dynamic CBCT reconstruction framework. The reconstruction framework reconstructs a dynamic sequence of CBCTs in a data-driven manner from a standard pre-treatment scan, without utilizing patient-specific knowledge. Meanwhile, a convolutional neural network-based motion encoder is jointly trained during the reconstruction to learn motion-related features relevant for real-time motion estimation, based on a single arbitrarily-angled x-ray projection. DREME was tested on digital phantom simulation and real patient studies.Main results:DREME accurately solved 3D respiration-induced anatomic motion in real time (~1.5 ms inference time for each x-ray projection). In the digital phantom study, it achieved an average lung tumor center-of-mass localization error of 1.2±0.9 mm (Mean±SD). In the patient study, it achieved a real-time tumor localization accuracy of 1.8±1.6 mm in the projection domain.Significance:DREME achieves CBCT and volumetric motion estimation in real time from a single x-ray projection at arbitrary angles, paving the way for future clinical applications in intra-fractional motion management. In addition, it can be used for dose tracking and treatment assessment, when combined with real-time dose calculation.

In this paper, we propose a systematic design framework to deal with the problem of limited CSIT feedback for MIMO-OFDM systems with correlated subcarriers. Based on the framework, we obtain the optimal transmission and CSI feedback strategies given the limited CSI feedback constraint. We propose a MIMO-OFDM design with combined adaptive power control and beam-forming framework for optimizing MIMO-OFDM link capacity with limited feedback in frequency selective fading channels. We derive a computationally efficient algorithm which exploits subcarrier correlation to search for the design of the optimal transmission and feedback strategy. We found that with a small number of bits for CSIT feedback, there is already significant capacity gain in the MIMO-OFDM systems

Objective.Real-time cone-beam computed tomography (CBCT) provides instantaneous visualization of patient anatomy for image guidance, motion tracking, and online treatment adaptation in radiotherapy. While many real-time imaging and motion tracking methods leveraged patient-specific prior information to alleviate under-sampling challenges and meet the temporal constraint (<500 ms), the prior information can be outdated and introduce biases, thus compromising the imaging and motion tracking accuracy. To address this challenge, we developed a frameworkdynamicreconstruction andmotionestimation (DREME) for real-time CBCT imaging and motion estimation, without relying on patient-specific prior knowledge.Approach.DREME incorporates a deep learning-based real-time CBCT imaging and motion estimation method into a dynamic CBCT reconstruction framework. The reconstruction framework reconstructs a dynamic sequence of CBCTs in a data-driven manner from a standard pre-treatment scan, without requiring patient-specific prior knowledge. Meanwhile, a convolutional neural network-based motion encoder is jointly trained during the reconstruction to learn motion-related features relevant for real-time motion estimation, based on a single arbitrarily-angled x-ray projection. DREME was tested on digital phantom simulations and real patient studies.Main Results.DREME accurately solved 3D respiration-induced anatomical motion in real time (∼1.5 ms inference time for each x-ray projection). For the digital phantom studies, it achieved an average lung tumor center-of-mass localization error of 1.2 ± 0.9 mm (Mean ± SD). For the patient studies, it achieved a real-time tumor localization accuracy of 1.6 ± 1.6 mm in the projection domain.Significance.DREME achieves CBCT and volumetric motion estimation in real time from a single x-ray projection at arbitrary angles, paving the way for future clinical applications in intra-fractional motion management. In addition, it can be used for dose tracking and treatment assessment, when combined with real-time dose calculation.

In this paper, we propose a systematic design framework to deal with the problem of partial CSIT feedback for MIMO-OFDM systems. Based on the framework, we obtain the optimal transmission and CSI feedback strategy given the limited CSI feedback constraint. We found that the design problem is very similar to the classical vector quantization problem with a modified distortion metric. We derive a computational efficient algorithm to search for the design of the optimal transmission and feedback strategy. We found that with a small number of bits for CSIT feedback, there is already significant capacity gain in the MIMO-OFDM systems. For illustration, we compare our proposed scheme with some regular adaptation schemes (designed for perfect CSIT). We found that systems designed for perfect CSIT performs poorly when the feedback channel can only support a limited number of bits. This justifies the importance of designing the system matched to the limited feedback constraint.

Purpose Nuclear emergency medical rescue is a critical component of the nuclear emergency response system, playing a vital role in safeguarding public life and health. To address the urgent need for rapid, wide-range radiation biodosimetry in nuclear emergency scenarios, this study utilized female C57BL/6J mice model to develop a machine learning (ML) framework for radiation-responsive biomarker screening and dose reconstruction across a broad dose range (0–12 Gy), laying a foundational preclinical basis for future translational research in human biodosimetry. Materials and methods The blood sample of mice was collected at 24 hours and seven days post-irradiation. The gene expression was evaluated by transcriptomic sequencing. Further, differential expression analysis, Spearman’s correlation filtering and Boruta algorithm were sequentially employed for screening radiation biomarkers. The stacking model integrating multiple ML algorithm was established for dose reconstruction. The gene expression was ultimately validated by more practical qRT-PCR method. Results Spearman’s correlation filtering and Boruta algorithm was employed to identify 172 highly robust biomarkers from an initial pool of 25,654 genes. By utilizing a stacked ensemble ML approach, high-accuracy dose reconstruction was achieved across a broad range of 0–12 Gy, with an R 2 of 0.952 and an RMSE of 0.938 Gy, significantly outperforming conventional regression analysis and individual ML models. Further refinement reduced the gene panel to just 15 key markers while preserving reconstruction accuracy comparable to the full 172-gene model. Experimental validation via qRT-PCR confirmed the reliability of these biomarkers, demonstrating the framework’s potential for translation into a field-deployable diagnostic platform for radiation exposure assessment. Conclusions We established ML framework that incorporates a multi-stage biomarker screening strategy and a stacking ML mode, to achieve rapid and accurate dose reconstruction across a wide dose range. This methodology provides a novel technical solution for nuclear emergency medical response.

Average speed information, which is essential for routing applications, is often missing in the freely available OpenStreetMap (OSM) road network. In this contribution, we propose an estimation framework, including different machine learning (ML) models that estimate rural roads’ average speed based on current road information in OSM. We rely on three datasets covering two regions in Chile and Australia. Google Directions API data serves as reference data. An appropriate estimation framework is presented, which involves supervised ML models, unsupervised clustering, and dimensionality reduction to generate new input features. The regression performance of each model with different input feature modes is evaluated on each dataset. The best performing model results in a coefficient of determination R2=80.43%, which is significantly better than previous approaches relying on domain-knowledge. Overall, the potential of the ML-based estimation framework to estimate the average speed with OSM road network data is demonstrated. This ML-based approach is data-driven and does not require any domain knowledge. In the future, we intend to focus on the generalization ability of the estimation framework concerning its application in different regions worldwide. The implementation of our estimation framework for an exemplary dataset is provided on GitHub.

Evaluation of evapotranspiration for exorheic basins in China using an improved estimate of terrestrial water storage change

PurposeTo enable rigid body motion‐tolerant parallel volumetric magnetic resonance imaging by retrospective head motion correction on a variety of spatiotemporal scales and imaging sequences.Theory and methodsTolerance against rigid body motion is based on distributed and incoherent sampling orders for boosting a joint retrospective motion estimation and reconstruction framework. Motion resilience stems from the encoding redundancy in the data, as generally provided by the coil array. Hence, it does not require external sensors, navigators or training data, so the methodology is readily applicable to sequences using 3D encodings.ResultsSimulations are performed showing full inter‐shot corrections for usual levels of in vivo motion, large number of shots, standard levels of noise and moderate acceleration factors. Feasibility of inter‐ and intra‐shot corrections is shown under controlled motion in vivo. Practical efficacy is illustrated by high‐quality results in most corrupted of 208 volumes from a series of 26 clinical pediatric examinations collected using standard protocols.ConclusionsThe proposed framework addresses the rigid motion problem in volumetric anatomical brain scans with sufficient encoding redundancy which has enabled reliable pediatric examinations without sedation.

Effective communication across languages remains a critical challenge, particularly in low-resource settings where conventional machine translation approaches falter due to sparse data and limited quality feedback. This paper presents a holistic framework to enhance reinforcement learning (RL) based machine translation systems tailored for such environments. We address the trifecta of challenges: sparse feedback on translation quality, ethical implications in algorithmic decision-making, and the imperative to adapt models to nuanced linguistic domains. This approach integrates advanced techniques in sparse reward handling, ensuring RL models learn efficiently despite limited feedback. Ethical considerations drive our methodology, emphasizing fairness, bias mitigation, and cultural sensitivity to uphold ethical standards in AI-driven translations. Additionally, domain-specific adaptation strategies are explored to tailor models to diverse linguistic contexts, from technical jargon to colloquialisms, enhancing translation accuracy and relevance. Through a rigorous experimental framework, including evaluation metrics like BLEU score and user feedback, we demonstrate substantial improvements in translation quality and ethical compliance compared to traditional methods. This research contributes to the evolution of robust, inclusive translation technologies pivotal for fostering global understanding and equitable access to information. This paper not only addresses current challenges but also sets a precedent for future research in AI ethics and machine learning applications, advocating for responsible innovation in crosscultural communication technologies

The accurate estimation and prediction of internal states in lithium-ion (Li-Ion) batteries, such as State of Charge (SoC) and Remaining Useful Life (RUL), are vital for optimizing battery performance, safety, and longevity in electric vehicles and other applications. This paper presents a unified, modular, and extensible machine learning (ML) framework designed to address the heterogeneity and complexity of battery state prediction tasks. The proposed framework supports flexible configurations across multiple dimensions, including feature engineering, model selection, and training/testing strategies. It integrates standardized data processing pipelines with a diverse set of ML models, such as a long short-term memory neural network (LSTM), a convolutional neural network (CNN), a feedforward neural network (FFNN), automated machine learning (AutoML), and classical regressors, while accommodating heterogeneous datasets. The framework’s applicability is demonstrated through five distinct use cases involving SoC estimation and RUL prediction using real-world and benchmark datasets. Experimental results highlight the framework’s adaptability, methodological transparency, and robust predictive performance across various battery chemistries, usage profiles, and degradation conditions. This work contributes to a standardized approach that facilitates the reproducibility, comparability, and practical deployment of ML-based battery analytics.

Spiking camera, a novel retina-inspired vision sensor, has shown its great potential for capturing high-speed dynamic scenes with a sampling rate of 40,000 Hz. The spiking camera abandons the concept of exposure window, with each of its photosensitive units continuously capturing photons and firing spikes asynchronously. However, the special sampling mechanism prevents the frame-based algorithm from being used to spiking camera. It remains to be a challenge to reconstruct dynamic scenes and perform common computer vision tasks for spiking camera. In this paper, we propose a self-supervised joint learning framework for optical flow estimation and reconstruction of spiking camera. The framework reconstructs clean frame-based spiking representations in a self-supervised manner, and then uses them to train the optical flow networks. We also propose an optical flow based inverse rendering process to achieve self-supervision by minimizing the difference with respect to the original spiking temporal aggregation image. The experimental results demonstrate that our method bridges the gap between synthetic and real-world scenes and achieves desired results in real-world scenarios. To the best of our knowledge, this is the first attempt to jointly reconstruct dynamic scenes and estimate optical flow for spiking camera from a self-supervised learning perspective.

We propose Deep Learning (DL) as a framework for simultaneous waveform estimation and image reconstruction in passive synthetic aperture radar (SAR). We consider a passive imaging scenario where the scene is illuminated by source of opportunity with known position, but transmitting an unknown waveform. Image reconstruction is then posed as a machine learning task for which a recurrent neural network (RNN) can be constructed by unfolding the iterations of a proximal gradient descent algorithm. We separate the unknown waveform from the known components of the forward model and parameterize the network operator by the waveform coefficients. The image estimate generated by the RNN is then mapped back to the data space by a linear decoder stage, so that the waveform coefficients are refined in an unsupervised manner using passive SAR measurements as training data. As a result, network becomes a recurrent auto-encoder. In this architecture, forward propagation recovers scene reflectivity, and backpropagation solves for waveform coefficients. The non-convex training routine is initialized using the partially known SAR forward model, which helps ensure that the optimization problem converges to a satisfactory optimal point.

Accurate food volume estimation is crucial for medical nutrition management and health monitoring applications. However, existing methods for estimating food volume are often constrained by the use of monocular data. They typically rely on specialised hardware such as 3D scanners, gather sensor-specific information such as depth data, or depend on camera calibration with reference objects. In this paper, we present VolE, a novel framework that leverages mobile device-driven 3D reconstruction to estimate food volume. VolE captures images and camera locations in free motion to generate precise 3D models, thanks to AR-capable mobile devices. To achieve real-world measurement, VolE is a reference- and depth-free framework that leverages food video segmentation for food mask generation. We also introduce a new food dataset encompassing the challenging scenarios absent in the previous benchmarks. Our experiments demonstrate that VolE outperforms the existing volume estimation techniques across multiple datasets by achieving 2.22% MAPE, highlighting its superior performance in food volume estimation. The source code is available at https://umairharon.github.io/VolE.

Almost every link in the chain of exploration seismology methods used to process recorded data has been affected by Green’s theorem. Among the seismic processes that can be related to, and/or have benefited from, Green’s theorem are wavelet estimation, multiple elimination, regularization, redatuming, imaging, deghosting, and interferometry. This tutorial on various seismic exploration methods derived from Green’s theorem emphasizes seismic data reconstruction (including regularization and redatuming) and its relationship to interferometry as well as to wavelet estimation and wavefield separation. The last decade has witnessed ever-increasing attention within the energy industry and its concomitant representation in the published literature to methods dealing with wavefield reconstruction through in-terferometry or virtual-source techniques. The attention has re- newed interest in Green’s theorem because all different ap-proaches to interferometry can be derived from it. This tutorial provides a derivation and explication of the limitations of interferometric techniques (when interferometry is used to process measured data from marine surface seismic experiments with controlled sources) as approximations to Green’s theorem. This tutorial provides a definite statement of the comprehensive framework given by Green’s theorem to wavefield reconstruction and shows how different techniques are directly understood as specific mathematical forms and/or approximations to the theorem. The use of approximations can have shortcomings and create artifacts. These artifacts and errors are also analyzed and explained. All methods discussed in this tutorial recognize their foundation on Green’s theorem and have a secure mathematical-physics cornerstone to recognize the assumptions behind distinct approximate solutions and to guide the search for more accurate, effective techniques.

Physics-constrained deep inverse framework for noisy Ising spin reconstruction and state estimation