Search

Simulation of high-definition binaural room impulse responses using conventional approaches involves a significant amount of computational resources, resulting in high computational time, making these approaches incapable of performing real-time high quality acoustic virtual reality. This research implemented a methodology for the rapid impulse response generation using the position of a moving listener inside a fixed sound field. The rapid generation of the impulse response is performed using its representative compressed dimension, with a smaller dimension than the original impulse response, learned by variational autoencoders and long short-term memory neural networks. First, the methodology selects a representative number of impulse responses covering the area of interest using a reliable room acoustic simulator. Second, it generates a dataset with sufficient impulse responses uniformly distributed through a data augmentation approach using a modified bilinear interpolation from the impulse responses previously simulated. Third, it applies an unsupervised model to positionally cluster the impulse responses to reduce the variability of the impulse responses in the given environment. Fourth, it splits the impulse response into time segments and generates a dataset per segment and cluster. Fifth, it trains a variational autocoder with a long short-term memory neural network model for each time segment cluster of impulse responses to infer the correspondent compressed impulse response part. In summary, the impulse response is generated by assigning the current listener position to the corresponding cluster and executing the decoders of the variational autoencoders with long short-term memory, trained previously. The findings are encouraging; the normalized mean absolute error of the impulse responses gathered by the interpolator and the impulse responses generated by the proposed model is less than 15% in the 88% of impulse responses reserved for testing. Moreover, the average of the absolute error of the interaural cross-correlation coefficient between the impulse responses obtained from the simulator and the proposed model is around 16%, implying that most acoustic characteristics of the real impulse are preserved. In addition, the computational time for generating the impulse response segment of 300 ms is approximately 65 ms, which is almost haft than the total system latency for a realistic auralization, 112 ms.

Drug discovery is a costly and time-consuming process, especially because of the significant expenses connected with the high percentage of clinical trial failures. As such, there is a need for new paradigms enabling the optimization of the various stages, from hit identification to market approval. The upsurge in the use of artificial intelligence (AI) technologies and the advent of deep learning (DL) demonstrated a lot of promise in rethinking and redesigning the traditional pipelines in drug discovery, including de novo molecular design. In this regard, generative models have greatly impacted the de novo design of molecules with desired properties and are being increasingly integrated into real world drug discovery campaigns. Herein, we will briefly appraise recent case studies utilizing generative models for chemical structure generation in the area of anticancer drug discovery. Finally, we will analyze current challenges and limitations as well as the possible strategies to overcome them, outlining potential future directions to advance this exciting field.

The ability of deep learning to process vast data and uncover concealed malicious patterns has spurred the adoption of deep learning methods within the cybersecurity domain. Nonetheless, a notable hurdle confronting cybersecurity researchers today is the acquisition of a sufficiently large dataset to effectively train deep learning models. Privacy and security concerns associated with using real-world organization data have made cybersecurity researchers seek alternative strategies, notably focusing on generating synthetic data. Generative adversarial networks (GANs) have emerged as a prominent solution, lauded for their capacity to generate synthetic data spanning diverse domains. Despite their widespread use, the efficacy of GANs in generating realistic cyberattack data remains a subject requiring thorough investigation. Moreover, the proficiency of deep learning models trained on such synthetic data to accurately discern real-world attacks and anomalies poses an additional challenge that demands exploration. This paper delves into the essential aspects of generative learning, scrutinizing their data generation capabilities, and conducts a comprehensive review to address the above questions. Through this exploration, we aim to shed light on the potential of synthetic data in fortifying deep learning models for robust cybersecurity applications. [ABSTRACT FROM AUTHOR]

Purpose: Due to various physical degradation factors and limited counts received, PET image quality needs further improvements. The denoising diffusion probabilistic model (DDPM) was a distribution learning-based model, which tried to transform a normal distribution into a specific data distribution based on iterative refinements. In this work, we proposed and evaluated different DDPM-based methods for PET image denoising. Methods: Under the DDPM framework, one way to perform PET image denoising was to provide the PET image and/or the prior image as the input. Another way was to supply the prior image as the network input with the PET image included in the refinement steps, which could fit for scenarios of different noise levels. 150 brain [ 18 F]FDG datasets and 140 brain [ 18 F]MK-6240 (imaging neurofibrillary tangles deposition) datasets were utilized to evaluate the proposed DDPM-based methods. Results: Quantification showed that the DDPM-based frameworks with PET information included generated better results than the nonlocal mean, Unet and generative adversarial network (GAN)-based denoising methods. Adding additional MR prior in the model helped achieved better performance and further reduced the uncertainty during image denoising. Solely relying on MR prior while ignoring the PET information resulted in large bias. Regional and surface quantification showed that employing MR prior as the network input while embedding PET image as a data-consistency constraint during inference achieved the best performance. Conclusion: DDPM-based PET image denoising is a flexible framework, which can efficiently utilize prior information and achieve better performance than the nonlocal mean, Unet and GAN-based denoising methods. [ABSTRACT FROM AUTHOR]

Deep generative models with latent variables have been used lately to learn joint representations and generative processes from multi-modal data, which depict an object from different viewpoints. These two learning mechanisms can, however, conflict with each other and representations can fail to embed information on the data modalities. This research studies the realistic scenario in which all modalities and class labels are available for model training, e.g. images or handwriting, but where some modalities and labels required for downstream tasks are missing, e.g. text or annotations. We show, in this scenario, that the variational lower bound limits mutual information between joint representations and missing modalities. We, to counteract these problems, introduce a novel conditional multi-modal discriminative model that uses an informative prior distribution and optimizes a likelihood-free objective function that maximizes mutual information between joint representations and missing modalities. Extensive experimentation demonstrates the benefits of our proposed model, empirical results show that our model achieves state-of-the-art results in representative problems such as downstream classification, acoustic inversion, and image and annotation generation. [ABSTRACT FROM AUTHOR]

Privacy concerns in the digital era highlight the sensitivity of facial images, revealing personal information beyond the identity. The current face deidentification techniques aiming to solve the issue often suffer from limitations. They either focus on narrow facial areas, compromise the image naturalness, lack flexibility in privacy levels, or offer suboptimal tradeoffs between the identity protection, image quality, and utility preservation. The paper presents a novel controllable face deidentification method that leverages StyleGAN2 and auxiliary classification approaches addressing these shortcomings. Validated across four datasets and compared to seven competitors, the method ensures a significant identity protection, preserves the data utility, maintains the diversity among deidentified faces, and demonstrates a promising performance. [ABSTRACT FROM AUTHOR]

Multitask learning (MTL) and data augmentation are becoming increasingly popular in natural language processing (NLP). These techniques are particularly useful when data are scarce. In MTL, knowledge learned from one task is applied to another. To address data scarcity, data augmentation facilitates by providing additional synthetic data during model training. In NLP, the bidirectional encoder representations from transformers (BERT) model is the default candidate for various tasks. MTL and data augmentation using BERT have yielded promising results. However, a detailed study regarding the effect of using MTL in different layers of BERT and the benefit of data augmentation in these configurations has not been conducted. In this study, we investigate the use of MTL and data augmentation from generative models, specifically for category classification, sentiment classification, and aspect-opinion sequence-labeling using BERT. The layers of BERT are categorized into top, middle, and bottom layers, which are frozen, shared, or unshared. Experiments are conducted to identify the optimal layer configuration for improved performance compared with that of single-task learning. Generative models are used to generate augmented data, and experiments are performed to reveal their effectiveness. The results indicate the effectiveness of the MTL configuration compared with single-task learning as well as the effectiveness of data augmentation using generative models for classification tasks. [ABSTRACT FROM AUTHOR]

Detecting conditional independencies plays a key role in several statistical and machine learning tasks, especially in causal discovery algorithms, yet it remains a highly challenging problem due to dimensionality and complex relationships presented in data. In this study, we introduce LCIT (Latent representation-based Conditional Independence Test)—a novel method for conditional independence testing based on representation learning. Our main contribution involves a hypothesis testing framework in which to test for the independence between X and Y given Z, we first learn to infer the latent representations of target variables X and Y that contain no information about the conditioning variable Z. The latent variables are then investigated for any significant remaining dependencies, which can be performed using a conventional correlation test. Moreover, LCIT can also handle discrete and mixed-type data in general by converting discrete variables into the continuous domain via variational dequantization. The empirical evaluations show that LCIT outperforms several state-of-the-art baselines consistently under different evaluation metrics, and is able to adapt really well to both nonlinear, high-dimensional, and mixed data settings on a diverse collection of synthetic and real data sets. [ABSTRACT FROM AUTHOR]

Data imputation is prevalent in crowdsensing, especially for Internet of Things (IoT) devices. On the one hand, data collected from sensors will inevitably be affected or damaged by unpredictability. On the other hand, extending the active time of sensor networks has urgently aspired environmental monitoring. Using neural networks to design a data imputation algorithm can take advantage of the prior information stored in the models. This paper proposes a preprocessing algorithm to extract a subset for training a neural network on an IoT dataset, including time window determination, sensor aggregation, sensor exclusion and data frame shape selection. Moreover, we propose a data imputation algorithm using deep compressed sensing with generative models. It explores novel representation matrices and can impute data in the case of a high missing ratio situation. Finally, we test our subset extraction algorithm and data imputation algorithm on the EPFL SensorScope dataset, respectively, and they effectively improve the accuracy and robustness even with extreme data loss. [ABSTRACT FROM AUTHOR]

Deep neural network approaches to inverse imaging problems have produced impressive results in the last few years. In this survey paper, we consider the use of generative models in a variational regularisation approach to inverse problems. The considered regularisers penalise images that are far from the range of a generative model that has learned to produce images similar to a training dataset. We name this family generative regularisers. The success of generative regularisers depends on the quality of the generative model and so we propose a set of desired criteria to assess generative models and guide future research. In our numerical experiments, we evaluate three common generative models, autoencoders, variational autoencoders and generative adversarial networks, against our desired criteria. We also test three different generative regularisers on the inverse problems of deblurring, deconvolution, and tomography. We show that restricting solutions of the inverse problem to lie exactly in the range of a generative model can give good results but that allowing small deviations from the range of the generator produces more consistent results. Finally, we discuss future directions and open problems in the field. [ABSTRACT FROM AUTHOR]

Data augmentation is a promising technique in improving exploration and convergence speed in deep reinforcement learning methodologies. In this work, we propose a data augmentation framework based on generative models for creating completely novel states and increasing diversity. For this purpose, a diffusion model is used to generate artificial states (learning the distribution of original, collected states), while an additional model is trained to predict the action executed between two consecutive states. These models are combined to create synthetic data for cases of high and low immediate rewards, which are encountered less frequently during the agent's interaction with the environment. During the training process, the synthetic samples are mixed with the actually observed data in order to speed up agent learning. The proposed methodology is tested on the Atari 2600 framework, producing realistic and diverse synthetic data which improve training in most cases. Specifically, the agent is evaluated on three heterogeneous games, achieving a reward increase of up to 31%, although the results indicate performance variance among the different environments. The augmentation models are independent of the learning process and can be integrated to different algorithms, as well as different environments, with slight adaptations. [ABSTRACT FROM AUTHOR]

With the rapid development of digital technologies and their implementation in everyday life, the topic of artificial intelligence (AI) is increasingly being discussed. This new to many technologies is constantly surprising with its capabilities, for example in the field of creation. Generative models, which are already capable of creating quite original visual, textual or musical content, are rapidly gaining attention and finding applications in a few sectors. The use of generative AI is also expanding significantly among students, for whom it promotes learning and creativity. However, this may pose a global problem in terms of ethics and copyright protection. The aim of this paper is to assess whether there is sufficient awareness of this topic at different school levels and what the practical experience of using AI among the younger generation is, based on an analysis of students' perceptions of AI. [ABSTRACT FROM AUTHOR]

In recent years, generative transformers have become increasingly prevalent in the field of artificial intelligence, especially within the scope of natural language processing. This paper provides a comprehensive overview of these models, beginning with the foundational theories introduced by Alan Turing and extending to contemporary generative transformer architectures. The manuscript serves as a review, historical account, and tutorial, aiming to offer a thorough understanding of the models' importance, underlying principles, and wide-ranging applications. The tutorial section includes a practical guide for constructing a basic generative transformer model. Additionally, the paper addresses the challenges, ethical implications, and future directions in the study of generative models. [ABSTRACT FROM AUTHOR]

In medical imaging, deep learning models serve as invaluable tools for expediting diagnoses and aiding specialized medical professionals in making clinical decisions. However, effectively training deep learning models typically necessitates substantial quantities of high-quality data, a resource often lacking in numerous medical imaging scenarios. One way to overcome this deficiency is to artificially generate such images. Therefore, in this comparative study we train five generative models to artificially increase the amount of available data in such a scenario. This synthetic data approach is evaluated on a a downstream classification task, predicting four causes for pneumonia as well as healthy cases on 1082 chest X-ray images. Quantitative and medical assessments show that a Generative Adversarial Network (GAN)-based approach significantly outperforms more recent diffusion-based approaches on this limited dataset with better image quality and pathological plausibility. We show that better image quality surprisingly does not translate to improved classification performance by evaluating five different classification models and varying the amount of additional training data. Class-specific metrics like precision, recall, and F1-score show a substantial improvement by using synthetic images, emphasizing the data rebalancing effect of less frequent classes. However, overall performance does not improve for most models and configurations, except for a DreamBooth approach which shows a +0.52 improvement in overall accuracy. The large variance of performance impact in this study suggests a careful consideration of utilizing generative models for limited data scenarios, especially with an unexpected negative correlation between image quality and downstream classification improvement. [ABSTRACT FROM AUTHOR]

Humans are able to quickly adapt to new situations, learn effectively with limited data, and create unique combinations of basic concepts. In contrast, generalizing out-of-distribution (OOD) data and achieving combinatorial generalizations are fundamental challenges for machine learning models. Moreover, obtaining high-quality labeled examples can be very time-consuming and expensive, particularly when specialized skills are required for labeling. To address these issues, we propose BtVAE, a method that utilizes conditional VAE models to achieve combinatorial generalization in certain scenarios and consequently to generate out-of-distribution (OOD) data in a semi-supervised manner. Unlike previous approaches that use new factors of variation during testing, our method uses only existing attributes from the training data but in ways that were not seen during training (e.g., small objects of a specific shape during training and large objects of the same shape during testing). [ABSTRACT FROM AUTHOR]

Denoising Diffusion models are gaining increasing popularity in the field of generative modeling for several reasons, including the simple and stable training, the excellent generative quality, and the solid probabilistic foundation. In this article, we address the problem of embedding an image into the latent space of Denoising Diffusion Models, that is finding a suitable "noisy" image whose denoising results in the original image. We particularly focus on Denoising Diffusion Implicit Models due to the deterministic nature of their reverse diffusion process. As a side result of our investigation, we gain a deeper insight into the structure of the latent space of diffusion models, opening interesting perspectives on its exploration, the definition of semantic trajectories, and the manipulation/conditioning of encodings for editing purposes. A particularly interesting property highlighted by our research, which is also characteristic of this class of generative models, is the independence of the latent representation from the networks implementing the reverse diffusion process. In other words, a common seed passed to different networks (each trained on the same dataset), eventually results in identical images. [ABSTRACT FROM AUTHOR]

Generative Adversarial Networks are powerful generative models that are used in different areas and with multiple applications. However, this type of model has a training problem called mode collapse. This problem causes the generator to not learn the complete distribution of the data with which it is trained. To force the network to learn the entire data distribution, MSSGAN is introduced. This model has multiple generators and distributes the training data in multiple subspaces, where each generator is enforced to learn only one of the groups with the help of a classifier. We demonstrate that our model performs better on the FID and Sample Distribution metrics compared to previous models to avoid mode collapse. Experimental results show how each of the generators learns different information and, in turn, generates satisfactory quality samples. [ABSTRACT FROM AUTHOR]

In this study, we investigate the application of generative models to assist artificial agents, such as delivery drones or service robots, in visualising unfamiliar destinations solely based on textual descriptions. We explore the use of generative models, such as Stable Diffusion, and embedding representations, such as CLIP and VisualBERT, to compare generated images obtained from textual descriptions of target scenes with images of those scenes. Our research encompasses three key strategies: image generation, text generation, and text enhancement, the latter involving tools such as ChatGPT to create concise textual descriptions for evaluation. The findings of this study contribute to an understanding of the impact of combining generative tools with multi-modal embedding representations to enhance the artificial agent's ability to recognise unknown scenes. Consequently, we assert that this research holds broad applications, particularly in drone parcel delivery, where an aerial robot can employ text descriptions to identify a destination. Furthermore, this concept can also be applied to other service robots tasked with delivering to unfamiliar locations, relying exclusively on user-provided textual descriptions. [ABSTRACT FROM AUTHOR]

With the recent advancements in the field of diffusion generative models, it has been shown that defining the generative process in the latent space of a powerful pretrained autoencoder can offer substantial advantages. This approach, by abstracting away imperceptible image details and introducing substantial spatial compression, renders the learning of the generative process more manageable while significantly reducing computational and memory demands. In this work, we propose to replace autoencoder coding with a model-based coding scheme based on traditional lossy image compression techniques; this choice not only further diminishes computational expenses but also allows us to probe the boundaries of latent-space image generation. Our objectives culminate in the proposal of a valuable approximation for training continuous diffusion models within a discrete space, accompanied by enhancements to the generative model for categorical values. Beyond the good results obtained for the problem at hand, we believe that the proposed work holds promise for enhancing the adaptability of generative diffusion models across diverse data types beyond the realm of imagery. [ABSTRACT FROM AUTHOR]

In this paper, we introduce a cutting-edge system that leverages state-of-the-art deep learning methodologies to generate high-quality synthetic thermal face images. Our unique approach integrates a thermally fine-tuned Stable Diffusion Model with a Vision Transformer (ViT) classifier, augmented by a Prompt Designer and Prompt Database for precise image generation control. Through rigorous testing across various scenarios, the system demonstrates its capability in producing accurate and superior-quality thermal images. A key contribution of our work is the development of a synthetic thermal face image database, offering practical utility for training thermal detection models. The efficacy of our synthetic images was validated using a facial detection model, achieving results comparable to real thermal face images. Specifically, a detector fine-tuned with real thermal images achieved a 97% accuracy rate when tested with our synthetic images, while a detector trained exclusively on our synthetic data achieved an accuracy of 98%. This research marks a significant advancement in thermal image synthesis, paving the way for its broader application in diverse real-world scenarios. [ABSTRACT FROM AUTHOR]

Copyright of Methaodos: Social Science Journal / Methaodos: Revista de Ciencias Sociales is the property of Departamento de Ciencias de la Comunicacion y Sociologia, Universidad Rey Juan Carlos and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

We propose an approach to symbolic regression based on a novel variational autoencoder for generating hierarchical structures, HVAE. It combines simple atomic units with shared weights to recursively encode and decode the individual nodes in the hierarchy. Encoding is performed bottom-up and decoding top-down. We empirically show that HVAE can be trained efficiently with small corpora of mathematical expressions and can accurately encode expressions into a smooth low-dimensional latent space. The latter can be efficiently explored with various optimization methods to address the task of symbolic regression. Indeed, random search through the latent space of HVAE performs better than random search through expressions generated by manually crafted probabilistic grammars for mathematical expressions. Finally, EDHiE system for symbolic regression, which applies an evolutionary algorithm to the latent space of HVAE, reconstructs equations from a standard symbolic regression benchmark better than a state-of-the-art system based on a similar combination of deep learning and evolutionary algorithms. [ABSTRACT FROM AUTHOR]