We discuss our insights into interpretable artificial-intelligence (AI) models, and how they are essential in the context of developing ethical AI systems, as well as data-driven solutions compliant with the Sustainable Development Goals (SDGs). We highlight the potential of extracting truly-interpretable models from deep-learning methods, for instance via symbolic models obtained through inductive biases, to ensure a sustainable development of AI.

In this work, we propose a two-stage autoencoder based compressor-decompressor framework for compressing malaria RBC cell image patches. We know that the medical images used for disease diagnosis are around multiple gigabytes size, which is quite huge. The proposed residual-based dual autoencoder network is trained to extract the unique features which are then used to reconstruct the original image through the decompressor module. The two latent space representations (first for the original image and second for the residual image) are used to rebuild the final original image. Color-SSIM has been exclusively used to check the quality of the chrominance part of the cell images after decompression. The empirical results indicate that the proposed work outperformed other neural network related compression technique for medical images by approximately 35%, 10% and 5% in PSNR, Color SSIM and MS-SSIM respectively. The algorithm exhibits a significant improvement in bit savings of 76%, 78%, 75% & 74% over JPEG-LS, JP2K-LM, CALIC and recent neural network approach respectively, making it a good compression-decompression technique.

This thesis investigates the controllability of deep learning-based, end-to-end, generative dialogue systems in both task-oriented and chit-chat scenarios. In particular, we study the different aspects of controlling generative dialogue systems, including controlling styles and topics and continuously adding and combining dialogue skills. In the three decades since the first dialogue system was commercialized, the basic architecture of such systems has remained substantially unchanged, consisting of four pipelined basic components, namely, natural language understanding (NLU), dialogue state tracking (DST), a dialogue manager (DM) and natural language generation (NLG). The dialogue manager, which is the critical component of the modularized system, controls the response content and style. This module is usually programmed by rules and is designed to be highly controllable and easily extendable. With the emergence of powerful "deep learning" architectures, end-to-end generative dialogue systems have been proposed to optimize overall system performance and simplify training. However, these systems cannot be easily controlled and extended as the modularized dialogue manager can. This is because a single neural system is used, which is usually a large pre-trained language model (e.g., GPT-2), and thus it is hard to surgically change desirable attributes (e.g., style, topics, etc.). More importantly, uncontrollable dialogue systems can generate offensive and even toxic responses. Therefore, in this thesis, we study controllable methods for end-to-end generative dialogue systems in task-oriented and chit-chat scenarios. Throughout the chapters, we describe 1) how to control the style and topics of chit-chat models, 2) how to continuously control and extend task-oriented dialogue systems, and 3) how to compose and control multi-skill dialogue models.

There is a mediocre content deluge coming to the internet the likes of which we have not seen. What if you could produce 10x the amount of content at at 10x cost savings, what would you do? Even if the content were mediocre would you still be tempted to take advantage of the ability to throw content against the well and see what sticks? What would that mean for websites, link farms, private blog networks, link builders, SEOs and search engine algorithms? What would it mean for quality, believable, original content?

In 2021, machine learning and deep learning had many amazing advances and important research papers may lead to breakthroughs in technology that get used by billions of people. The research in this field is developing very quickly and to help you monitor the progress here is the list of most important recent scientific research papers. Presented a technique for fair-minded gradient assessment in untolled calculation charts, called Persistent Evolution Strategies (PES). PES acquires inclinations from truncated unrolls, which speeds up streamlining by taking into consideration that frequent parameter updates while not experiencing truncation predisposition that influences many contending approaches. The researchers showed PES is extensively relevant, with tests exhibiting its application to an RNN-like task, support learning, etc. Showed tensor trains give an engaging estimate system to illustrative partial differential equations (PDEs): the mix of reformulations as far as in reverse stochastic differential equations and relapse type techniques in the tensor format hold the guarantee of utilizing latent low-rank designs empowering both pressure and effective calculation.

In recent years machines have learned to generate passable snippets of English, thanks to advances in artificial intelligence. Now they are moving on to other languages. Aleph Alpha, a startup in Heidelberg, Germany, has built one of the world's most powerful AI language models. Befitting the algorithm's European origins, it is fluent not just in English but also in German, French, Spanish, and Italian. The algorithm builds on recent advances in machine learning that have helped computers handle language with what sometimes seems like real understanding.

Landslides pose a serious risk to population, property, and environment in mountainous regions and even in flat areas worldwide. Landslides have caused massive casualties and significant losses and damage to property. In recent years, machine learning (ML) techniques, including deep learning methods, have increasingly been used to model complex landslides. Analyses so far have demonstrated promising predictive ability compared to traditional, deterministic solutions, and physical model testing. This Special Issue of Applied Sciences seeks to incorporate the latest developments in machine learning with respect to modeling and prediction of landslide susceptibility, including quantitative and qualitative assessments of the classification, volume (or area) and spatial distribution of landslides, as well as the velocity, intensity, and runout (and consequences) of existing or potential landsliding.

Note: All of the code is in the form of snippets and will not work when executed alone. The full code can be found on my Github repo. Reinforcement learning is the training of an agent to make decisions in an environment. An agent is deployed in an environment. At any given frame, the agent must use data from the environment to act.

Both Convolution and Deconvolution are used for image processing. The goal of the convolution layer is to transform an image into an abstract representation of output values. The convolution algorithm downsizes the image that means if the size of the input image is 6 x 6 then the image will be downsized to a 4 x 4 or 3 x 3. Deconvolution is a mathematical operation that reverses the process of the convolution layer. Now throw the output through the deconvolutional layer, and you get back the same input. The importance of Deconvolution is to regenerating the image from the convolution output layer, by doing this we can be sure that the convolution is extracting the correct features from the image.

Today, AI technology is showing its strengths in almost every industry and walks of life. From text generation, text summarization, chatbots, NLP is being used widely. One such paradigm is automatic code generation. An AI could be generating anything; hence the output space is unconstrained. A self-driving car is driven for 100 million miles to validate its safety, but tests cannot be written to monitor and cover an unconstrained space. One of the solutions to validate AI-generated content is to constrain the problem and convert it from abstract to realistic, and this can be accomplished by either validating the unconstrained algorithm using theoretical proofs or by using Monte-Carlo simulation methods. In this case, we use the latter approach to test/validate a statistically significant number of samples. This hypothesis of validating the AI-generated code is the main motive of this work and to know if AI-generated code is reliable, a metric model CGEMs is proposed. This is an extremely challenging task as programs can have different logic with different naming conventions, but the metrics must capture the structure and logic of the program. This is similar to the importance grammar carries in AI-based text generation, Q&A, translations, etc. The various metrics that are garnered in this work to support the evaluation of generated code are as follows: Compilation, NL description to logic conversion, number of edits needed, some of the commonly used static-code metrics and NLP metrics. These metrics are applied to 80 codes generated using OpenAI's GPT-3. Post which a Neural network is designed for binary classification (acceptable/not acceptable quality of the generated code). The inputs to this network are the values of the features obtained from the metrics. The model achieves a classification accuracy of 76.92% and an F1 score of 55.56%. XAI is augmented for model interpretability.