Diffusion Probabilistic models have been shown to generate state-of-the-art results on several competitive image synthesis benchmarks but lack a low-dimensional, interpretable latent space, and are slow at generation. On the other hand, Variational Autoencoders (VAEs) typically have access to a low-dimensional latent space but exhibit poor sample quality. Despite recent advances, VAEs usually require high-dimensional hierarchies of the latent codes to generate high-quality samples. We present DiffuseVAE, a novel generative framework that integrates VAE within a diffusion model framework, and leverage this to design a novel conditional parameterization for diffusion models. We show that the resulting model can improve upon the unconditional diffusion model in terms of sampling efficiency while also equipping diffusion models with the low-dimensional VAE inferred latent code. Furthermore, we show that the proposed model can generate high-resolution samples and exhibits synthesis quality comparable to state-of-the-art models on standard benchmarks. Lastly, we show that the proposed method can be used for controllable image synthesis and also exhibits out-of-the-box capabilities for downstream tasks like image super-resolution and denoising. For reproducibility, our source code is publicly available at https://github.com/kpandey008/DiffuseVAE. 1

The k-nearest-neighbor (kNN) decision rule is a simple and robust classifier for text categorization. The performance of kNN decision rule depends heavily upon the value of the neighborhood parameter k. The method categorize a test document even if the difference between the number of members of two competing categories is one. Hence, choice of k is crucial as different values of k can change the result of text categorization. Moreover, text categorization is a challenging task as the text data are generally sparse and high dimensional. Note that, assigning a document to a predefined category for an arbitrary value of k may not be accurate when there is no bound on the margin of majority voting. A method is thus proposed in spirit of the nearest-neighbor decision rule using a medoid-based weighting scheme to deal with these issues. The method puts more weightage on the training documents that are not only lie close to the test document but also lie close to the medoid of its corresponding category in decision making, unlike the standard nearest-neighbor algorithms that stress on the documents that are just close to the test document. The aim of the proposed classifier is to enrich the quality of decision making. The empirical results show that the proposed method performs better than different standard nearest-neighbor decision rules and support vector machine classifier using various well-known text collections in terms of macro- and micro-averaged f-measure.

<p>Presently, 17 out of 30 Indian cities are ranked worst in air quality around the globe due to high emissions of fine particulate matter, PM<sub>2.5</sub> (particles less than 2.5 µm diameter). These particles can reach deeper into the lungs and cause serious health problems, including cardiovascular obstructive pulmonary disease, lung cancer, stroke, and asthma. To take prompt actions towards mitigating and controlling the adverse effects of air pollution, it is important to monitor the ambient air quality regularly and at the neighbourhood level. However, the distribution of the regulatory central ambient air quality monitoring stations (CAAQMS) in India is sparse, and many states and cities lack any regulatory stationary monitors (RSMs). Conventional air quality monitoring techniques are inefficient and incapable of mapping PM<sub>2.5 </sub>at a sub-Km level. The heterogeneity of PM<sub>2.5 </sub>concentrations at large-scale and high spatial resolution has numerous applications in epidemiological studies, detecting hotspots within neighbourhoods and implementing policy interventions at local, regional and city levels. Therefore, an integrated monitoring framework is needed to fill gaps in the existing air quality measurements. This study proposes a tribrid approach of using the low-cost sensor (LCS) network to supplement the RSMs in generating more ground-truth PM<sub>2.5 </sub>concentrations along with high-resolution micro-satellite imageries (PlanetScope, ~3m/pixel) to estimate and generate the PM<sub>2.5 </sub>concentration maps at the sub-Km level (~500m by 500m). In the present study, an extensive LCS network of 70 nodes deployed at optimally selected locations within and around the boundaries of Lucknow city, Uttar Pradesh, India, along with six existing RSMs for one year (December 2021 onwards). It has increased monitoring ten folds at a moderate cost, covering remote urban and rural areas. The locations of these LCS and RSMs (76 nodes) have been used to precisely extract the daily (every day Dec 2021-2022) high-resolution satellite imageries by forming the area of interest (AOI) of size 224 by 224-pixel around the node while keeping the node in the middle of AOI. These imageries have been labelled with the ground truth PM<sub>2.5  </sub>values from the nodes with geographical location and meteorological parameters such as relative humidity, atmospheric temperature, and barometric pressure. These labelled data are then fed into a deep learning CNN-RT-RF (Convolutional neural network- random trees-random forest) joint model to predict PM<sub>2.5</sub> at sub-Km level, which provides RMSE~ 2.74 and 7.50 for training and test data, respectively. The study further compares model performance with existing datasets of Delhi and Beijing. The results show that the predicted PM<sub>2.5</sub> using satellite imagery shows a strong co-relation with LCS and RSMs network and thus can be used as a soft sensor for large-scale monitoring. This study is the first study to integrate LCS sensor data with microsatellite imagery, leveraging over costly, conventional methods using machine learning approaches.</p> <p> </p>

Normalizing flows provide an elegant method for obtaining tractable density estimates from distributions by using invertible transformations. The main challenge is to improve the expressivity of the models while keeping the invertibility constraints intact. We propose to do so via the incorporation of localized self-attention. However, conventional self-attention mechanisms don't satisfy the requirements to obtain invertible flows and can't be naively incorporated into normalizing flows. To address this, we introduce a novel approach called Attentive Contractive Flow (ACF) which utilizes a special category of flow-based generative models -contractive flows. We demonstrate that ACF can be introduced into a variety of state of the art flow models in a plug-and-play manner. This is demonstrated to not only improve the representation power of these models (improving on the bits per dim metric), but also to results in significantly faster convergence in training them. Qualitative results, including interpolations between test images, demonstrate that samples are more realistic and capture local correlations in the data well. We evaluate the results further by performing perturbation analysis using AWGN demonstrating that ACF models (especially the dotproduct variant) show better and more consistent resilience to additive noise.