Understanding human preferences is crucial for improving foundation models and building personalized AI systems. However, preferences are inherently diverse and complex, making it difficult for traditional reward models to capture their full range. While fine-grained preference data can help, collecting it is expensive and hard to scale. In this paper, we introduce Decomposed Reward Models (DRMs), a novel approach that extracts diverse human preferences from binary comparisons without requiring fine-grained annotations. Our key insight is to represent human preferences as vectors and analyze them using Principal Component Analysis (PCA). By constructing a dataset of embedding differences between preferred and rejected responses, DRMs identify orthogonal basis vectors that capture distinct aspects of preference. These decomposed rewards can be flexibly combined to align with different user needs, offering an interpretable and scalable alternative to traditional reward models. We demonstrate that DRMs effectively extract meaningful preference dimensions (e.g., helpfulness, safety, humor) and adapt to new users without additional training. Our results highlight DRMs as a powerful framework for personalized and interpretable LLM alignment.
It is difficult to identify optimal cut-off frequencies for filters used with the common spatial pattern (CSP) method in motor imagery (MI)-based brain-computer interfaces (BCIs). Most current studies choose filter cut-frequencies based on experience or intuition, resulting in sub-optimal use of MI-related spectral information in the electroencephalography (EEG). To improve information utilization, we propose a SincNet-based hybrid neural network (SHNN) for MI-based BCIs. First, raw EEG is segmented into different time windows and mapped into the CSP feature space. Then, SincNets are used as filter bank band-pass filters to automatically filter the data. Next, we used squeeze-and-excitation modules to learn a sparse representation of the filtered data. The resulting sparse data were fed into convolutional neural networks to learn deep feature representations. Finally, these deep features were fed into a gated recurrent unit module to seek sequential relations, and a fully connected layer was used for classification. We used the BCI competition IV datasets 2a and 2b to verify the effectiveness of our SHNN method. The mean classification accuracies (kappa values) of our SHNN method are 0.7426 (0.6648) on dataset 2a and 0.8349 (0.6697) on dataset 2b, respectively. The statistical test results demonstrate that our SHNN can significantly outperform other state-of-the-art methods on these datasets.
Abstract Purpose Current attenuation correction (AC) of myocardial perfusion (MP) positron emission tomography (PET) remains challenging in routine clinical practice due to the propagation of CT-based artifacts and potential mismatch between PET and CT. The goal of this work is to demonstrate the feasibility of directly generating attenuation-corrected PET (AC PET) images from non-attenuation-corrected PET (NAC PET) images in the reconstruction domain for [ 13 N]ammonia MP PET based on a generative adversarial network (GAN). Methods We recruited 60 patients who underwent rest [ 13 N]ammonia cardiac PET/CT examinations. One static frame and twenty-one dynamic frames were acquired for each patient with both NAC PET and CT-based AC (CTAC) PET images. Paired 3D static or dynamic NAC and CTAC PET images were used as network inputs and labels for static (S-DLAC) and dynamic (D-DLAC) MP PET, respectively. In addition, the pre-trained S-DLAC network was fine-tuned by 3D paired dynamic NAC and CTAC PET frames for then AC in the dynamic PET images (D-DLAC-FT). Qualitative and quantitative assessments were implemented using CTAC PET as reference. Results The proposed S-DLAC, D-DLAC and D-DLAC-FT methods were qualitatively and quantitatively consistent with clinical CTAC. The S-DLAC showed a higher correlation with the reference static CTAC (S-CTAC) as compared to static NAC. The estimated kinetic parameters and blood volume fraction images from D-DLAC and D-DLAC-FT methods showed comparable performances with the reference dynamic CTAC (D-CTAC). D-DLAC-FT was slightly better than D-DLAC in terms of various physical and clinical indices. Conclusion The proposed S-DLAC, D-DLAC and D-DLAC-FT methods reduced attenuation artifacts significantly and achieved comparable performance with clinical CTAC for static and dynamic cardiac PET. The use of transfer learning is effective for the dynamic MP PET AC purpose.
Dynamic positron emission tomography (PET) imaging usually suffers from high statistical noise due to low counts of the short frames. This study aims to improve the image quality of the short frames by utilizing information from other modality. We develop a deep learning-based joint filtering framework for simultaneously incorporating information from longer acquisition PET frames and high-resolution magnetic resonance (MR) images into the short frames. The network inputs are noisy PET images and corresponding MR images while the outputs are linear coefficients of spatially variant linear representation model. The composite of all dynamic frames is used as training label in each sample, and it is down-sampled to 1/10th of counts as the training input. L1-norm combined with two gradient-based regularizations constitute the loss function during training. Ten realistic dynamic PET/MR phantoms based on BrainWeb are used for pre-training and eleven clinical subjects from Alzheimer's Disease Neuroimaging Initiative further for fine-tuning. Simulation results show that the proposed method can reduce the statistical noise while preserving image details and achieve quantitative enhancements compared with Gaussian, guided filter, and convolutional neural network trained with the mean squared error. The clinical results perform better than others in terms of the mean activity and standard deviation. All of the results indicate that the proposed deep learning-based joint filtering framework is of great potential for dynamic PET image denoising.
Abstract With the rapid development of the network, network transmission encryption technologies such as SSL and SSH have emerged. Network traffic has grown exponentially, and transmission encryption has become an important means of protecting data security and privacy. However, encrypted data also brings hidden dangers that are not easily detectable to network security. Identifying the encrypted network traffic can effectively solve this problem. However, the current recognition probability is not high enough and the time delay caused by the recognition together makes it impossible to accurately detect and warn the network traffic. An encrypted network traffic recognition method based on deep learning is proposed. Experimental verification shows that the method is applied in the network. The accuracy of encrypted network traffic identification is 97.02%, which can meet actual needs.
The feature points extracted by the traditional ORB algorithm are not evenly distributed, redundant and have no scale invariance. To solve this problem, this paper improved the traditional ORB algorithm and proposed an optimized feature point extraction method. The image is divided into regions firstly. According to the total number of feature points to be extracted and the number of divided regions, the algorithm calculates the number of feature points to be extracted for each region, which solves the problem of feature point overlap and redundancy in the feature point extraction process. By constructing the image pyramid and extracting feature points on each layer, the problem that the feature points extracted by ORB algorithm do not have scale invariance is solved. The experimental results show that the feature points extracted by our algorithm are more uniform and reasonable without losing the accuracy of image matching, and the extraction speed is about 16% higher than that of the traditional ORB algorithm.
Abstract Background Low-dose ungated CT is commonly used for total-body PET attenuation and scatter correction (ASC). However, CT-based ASC (CT-ASC) is limited by radiation dose risks of CT examinations, propagation of CT-based artifacts and potential mismatches between PET and CT. We demonstrate the feasibility of direct ASC for multi-tracer total-body PET in the image domain. Methods Clinical uEXPLORER total-body PET/CT datasets of [ 18 F]FDG ( N = 52), [ 18 F]FAPI ( N = 46) and [ 68 Ga]FAPI ( N = 60) were retrospectively enrolled in this study. We developed an improved 3D conditional generative adversarial network (cGAN) to directly estimate attenuation and scatter-corrected PET images from non-attenuation and scatter-corrected (NASC) PET images. The feasibility of the proposed 3D cGAN-based ASC was validated using four training strategies: (1) Paired 3D NASC and CT-ASC PET images from three tracers were pooled into one centralized server (CZ-ASC). (2) Paired 3D NASC and CT-ASC PET images from each tracer were individually used (DL-ASC). (3) Paired NASC and CT-ASC PET images from one tracer ([ 18 F]FDG) were used to train the networks, while the other two tracers were used for testing without fine-tuning (NFT-ASC). (4) The pre-trained networks of (3) were fine-tuned with two other tracers individually (FT-ASC). We trained all networks in fivefold cross-validation. The performance of all ASC methods was evaluated by qualitative and quantitative metrics using CT-ASC as the reference. Results CZ-ASC, DL-ASC and FT-ASC showed comparable visual quality with CT-ASC for all tracers. CZ-ASC and DL-ASC resulted in a normalized mean absolute error (NMAE) of 8.51 ± 7.32% versus 7.36 ± 6.77% ( p < 0.05), outperforming NASC ( p < 0.0001) in [ 18 F]FDG dataset. CZ-ASC, FT-ASC and DL-ASC led to NMAE of 6.44 ± 7.02%, 6.55 ± 5.89%, and 7.25 ± 6.33% in [ 18 F]FAPI dataset, and NMAE of 5.53 ± 3.99%, 5.60 ± 4.02%, and 5.68 ± 4.12% in [ 68 Ga]FAPI dataset, respectively. CZ-ASC, FT-ASC and DL-ASC were superior to NASC ( p < 0.0001) and NFT-ASC ( p < 0.0001) in terms of NMAE results. Conclusions CZ-ASC, DL-ASC and FT-ASC demonstrated the feasibility of providing accurate and robust ASC for multi-tracer total-body PET, thereby reducing the radiation hazards to patients from redundant CT examinations. CZ-ASC and FT-ASC could outperform DL-ASC for cross-tracer total-body PET AC.
Image pairing is an important research task in the field of computer vision. And finding image pairs containing objects of the same category is the basis of many tasks such as tracking and person re-identification, etc., and it is also the focus of our research. Existing traditional methods and deep learning-based methods have some degree of defects in speed or accuracy. In this paper, we made improvements on the Siamese network [1] and proposed GetNet. The proposed method GetNet combines STN [2] and Siamese network to get the target area first and then perform subsequent processing. Experiments show that our method achieves competitive results in speed and accuracy.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)
