Accurate battery state-of-temperature estimation is essential for battery management systems (BMS) to ensure reliable operation, longevity, and safety. A fast, predictive model with an optimal grid search is needed to develop a data-driven battery model for future model-predictive control BMSs. This work presents a multiphysics-augmented feedforward neural network based on the feedforward multi-layer perceptron (MLP) to predict lithium-ion battery temperature (Tb) under driving conditions. A novelty in this work is the quality of the training data generated from an electrochemical NCA/Gr.SiOx battery model, validated against various driving data from Tesla Model Y experiments. The vehicle tests are performed under multiple battery self-cooling-heating modes (25–40 °C), SOC = 20–46%, and drive cycles (WLTC, FTP75, HWFET). The global sensitivity analysis and genetic algorithm are combined to optimize the cell model parameterizations, comprising geometric, transport/concentration, and kinetic parameters. The MLP grid search (hyperparameters) is optimized for the neural network model. The overall trends in the battery’s thermal states are predictable using the well-trained MLP model, with R2 values of 0.9–0.99. The true and predicted Tb maps as functions of cell geometric, transport, concentration, and kinetics parameters of the NCA/Gr.SiOx cell also match reasonably well. Due to its simple structure and ability to predict real-time battery thermal states, this work can serve as a guideline for designing data-driven battery models for future data-driven BMS applications.
Low-dose computed tomography (LDCT) is critical for minimizing radiation exposure, but it often leads to increased noise and reduced image quality. Traditional denoising methods, such as iterative optimization or supervised learning, often fail to preserve image quality. To address these challenges, we introduce PPORLD-EDNetLDCT, a reinforcement learning-based (RL) approach with Encoder–Decoder for LDCT. Our method utilizes a dynamic RL-based approach in which the Proximal Policy Optimization (PPO) algorithm is employed to optimize the denoising policy during training, guided by image quality feedback in a custom gym environment, while inference is performed using the trained fixed-parameter encoder–decoder model. The experimental results on the low dose CT image and projection dataset demonstrate that the proposed PPORLD-EDNetLDCT model outperforms traditional denoising techniques and other DL-based methods, achieving a peak signal-to-noise ratio of 41.87, a structural similarity index measure of 0.9814 and a root mean squared error of 0.00236. Moreover, in NIH-AAPM-Mayo Clinic Low Dose CT Challenge dataset our method achieved a PSNR of 41.52, SSIM of 0.9723 and RMSE of 0.0051. Furthermore, we validated the quality of denoising using a classification task in the COVID-19 LDCT dataset, where the images processed by our method improved the classification accuracy to 94%, achieving 4% higher accuracy compared to denoising without RL-based denoising.
Accurate land cover classification from satellite imagery is crucial in environmental monitoring and sustainable resource management. However, it remains challenging due to the complexity of natural landscapes, the visual similarity between classes, and the significant class imbalance in the available datasets. To address these issues, we propose a dual encoder architecture that independently extracts modality-specific features from optical and Synthetic Aperture Radar (SAR) imagery, which are then fused using a cross-modality attention-fusion module named Cross-modality Land cover segmentation with Attention and Imbalance-aware Reasoning-Enhanced Explanations (CLAIRE). This fusion mechanism highlights complementary spatial and textural features, enabling the network to better capture detailed and diverse land cover patterns. We incorporate a hybrid loss function that utilizes Weighted Focal Loss and Tversky Loss named RIFT (Rare-Instance Focal-Tversky) to address class imbalance and improve segmentation performance across underrepresented categories. Our model achieves competitive performance across multiple benchmarks: a mean Intersection over Union (mIoU) of 56.02% and Overall Accuracy (OA) of 84.56% on the WHU-OPT-SAR dataset; strong generalization with a mIoU of 59.89% and OA of 73.92% on the OpenEarthMap-SAR dataset; and remarkable robustness under cloud-obstructed conditions, achieving an mIoU of 86.86% and OA of 94.58% on the PIE-RGB-SAR dataset. Additionally, we introduce a metric-driven reasoning module generated by a Small Language Model (Phi-3), which generates sample-specific justifications for model predictions, thereby enhancing transparency and interpretability.
Recent advances in multimodal large language models (MLLMs) have delivered remarkable gains in multimodal understanding and reasoning, yet their prohibitive computational and memory costs remain a major barrier to real-world deployment. We present Mamba-based Multimodal Vision-Language Quantization (Mamba-MVQ), a compact and efficient multimodal inference framework that unifies the linear-complexity Mamba architecture with alignment-aware mixed-precision quantization. The proposed method assigns higher precision to semantically critical cross-modal modules, such as visual projection layers, while aggressively quantizing less sensitive components, including the linear layers in Mamba blocks, to maximize compression without severely compromising performance. To mitigate cross-modal semantic degradation introduced by quantization, we further introduce an alignment-preserving calibration objective based on vision-language contrastive learning. Combined with dynamic token merging, Mamba-MVQ reduces inference cost from both architectural and token-level perspectives. Extensive experiments on OpenCompass and MMBench show that Mamba-MVQ reduces model memory to 23.8% of the full-precision baseline with only a 1.7% average accuracy drop, demonstrating an effective trade-off between efficiency and multimodal capability.
