Accurate and timely regional weather prediction is vital for sectors dependent on weather-related decisions. Traditional prediction methods, based on atmospheric equations, often struggle with coarse temporal resolutions and inaccuracies. This paper presents a novel machine learning (ML) model, called MiMa (short for Micro-Macro), that integrates both near-surface observational data from Kentucky Mesonet stations (collected every five minutes, known as Micro data) and hourly atmospheric numerical outputs (termed as Macro data) for fine-resolution weather forecasting. The MiMa model employs an encoder-decoder transformer structure, with two encoders for processing multivariate data from both datasets and a decoder for forecasting weather variables over short time horizons. Each instance of the MiMa model, called a modelet, predicts the values of a specific weather parameter at an individual Mesonet station. The approach is extended with Re-MiMa modelets, which are designed to predict weather variables at ungauged locations by training on multivariate data from a few representative stations in a region, tagged with their elevations. Re-MiMa (short for Regional-MiMa) can provide highly accurate predictions across an entire region, even in areas without observational stations. Experimental results show that MiMa significantly outperforms current models, with Re-MiMa offering precise short-term forecasts for ungauged locations, marking a significant advancement in weather forecasting accuracy and applicability.

In order to address real-world problems, deep learning models are jointly trained on many classes. However, in the future, some classes may become restricted due to privacy/ethical concerns, and the restricted class knowledge has to be removed from the models that have been trained on them. The available data may also be limited due to privacy/ethical concerns, and re-training the model will not be possible. We propose a novel approach to address this problem without affecting the model's prediction power for the remaining classes. Our approach identifies the model parameters that are highly relevant to the restricted classes and removes the knowledge regarding the restricted classes from them using the limited available training data. Our approach is significantly faster and performs similar to the model re-trained on the complete data of the remaining classes.

GNW 9 million individuals are likely to be diagnosed with various types of cancer in the US. During the same year, around 0.6 million cancer-related deaths are anticipated to be reported in the aforementioned region. This, in turn, is expected to result in an increase of 70% in the global cancer burden, over the next two decades. Amidst the ever growing cancer burden, a number of strategies are being tested by researchers and industry players to help provide relief to the affected individuals. In recent years, artificial intelligence (AI) has emerged as a key enabler in improving the accuracy and speed of cancer diagnosis.

This paper is about the meaning of understanding in scientific and in artificial intelligent systems. We give a mathematical definition of the understanding, where, contrary to the common wisdom, we define the probability space on the input set, and we treat the transformation made by an intelligent actor not as a loss of information, but instead a reorganization of the information in the framework of a new coordinate system. We introduce, following the ideas of physical renormalization group, the notions of relevant and irrelevant parameters, and discuss, how the different AI tasks can be interpreted along these concepts, and how the process of learning can be described. We show, how scientific understanding fits into this framework, and demonstrate, what is the difference between a scientific task and pattern recognition. We also introduce a measure of relevance, which is useful for performing lossy compression.

In my last post, I discussed the state of confusion around deep learning and its abilities. Also, how even software programmers have a hard time understanding how deep learning enables machines to program themselves. In this post, I will try to explain probably the hardest to understand deep learning concept i.e. how deep learning machines program themselves without any human intervention. Since the advent of software programming, humans have been writing code to program the behavior of machines. In other words, the behavior of a machine only changes when the machine is reprogrammed by a human through new lines of code.

In my last post, I discussed the state of confusion around deep learning and its abilities. Also, how even software programmers have a hard time understanding how deep learning enables machines to program themselves. In this post, I will try to explain probably the hardest to understand deep learning concept i.e. how deep learning machines program themselves without any human intervention. Since the advent of software programming, humans have been writing code to program the behavior of machines. In other words, the behavior of a machine only changes when the machine is reprogrammed by a human through new lines of code.