Department of Biomedical Engineering and Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, NY, USA Abstract --Accurate sleep stage classification is essential for understanding sleep disorders and improving overall health. This study proposes a novel three-stage approach for sleep stage classification using ECG signals, offering a more accessible alternative to traditional methods that often rely on complex modalities like EEG. In Stages 1 and 2, we initialize the weights of two networks, which are then integrated in Stage 3 for comprehensive classification. In the first phase, we estimate key features using Feature Imitating Networks (FINs) to achieve higher accuracy and faster convergence. The second phase focuses on identifying the N1 sleep stage through the time-frequency representation of ECG signals. Finally, the third phase integrates models from the previous stages and employs a Kolmogorov-Arnold Network (KAN) to classify five distinct sleep stages. Additionally, data augmentation techniques, particularly SMOTE, are used in enhancing classification capabilities for underrepresented stages like N1. Our results demonstrate significant improvements in the classification performance, with an overall accuracy of 80.79% an overall kappa of 0.73. The model achieves specific accuracies of 86.70% for Wake, 60.36% for N1, 83.89% for N2, 84.85% for N3, and 87.16% for REM. This study emphasizes the importance of weight initialization and data augmentation in optimizing sleep stage classification with ECG signals. Keywords: Sleep Stage Classification, Electrocardiogram, Kolmogorov-Arnold Networks, Liquid Neural Networks. Sleep disorders are a persistent challenge throughout the human history.

The "quantum chandelier" that sits within a glass box in the chapel at Rensselaer Polytechnic Institute's campus in Troy, New York, is the symbolic centerpiece of an ambitious effort to turn upstate New York into an advanced technology center – what Silicon Valley is to social media or Cambridge, Massachusetts, is to biotech. The silver sci-fi object, named for interior gold lattices that suspend, cool and isolate its processor, is the heart of a "quantum computing system" that could herald a new age of computing. It's the centerpiece of the dream Curtis Priem, a co-founder of Nvidia, the 2.8tn artificial intelligence hardware and software company, has of turning Rensselaer, or RPI, into an advanced computing hub and refashioning this area of upstate New York into a new Silicon Valley. Priem has invested a sizable chunk of his fortune into building the Curtis Priem Quantum Constellation – a workshop for RPI students' vision of a quantum computing future. Just as his partners at Nvidia, where he was the company's first chief technology officer, gave him the freedom to imagine graphics chip architecture that powers the AI revolution, he hoped his investment will spark a new era of computer-powered innovation in the region.

Principal Components Analysis (PCA) is often used as a feature extraction procedure.

This work presents novel algorithms for learning Bayesian networks of bounded treewidth. Both exact and approximate methods are developed. The exact method combines mixed integer linear programming formulations for structure learning and treewidth computation. The approximate method consists in sampling k-trees (maximal graphs of treewidth k), and subsequently selecting, exactly or approximately, the best structure whose moral graph is a subgraph of that k-tree. The approaches are empirically compared to each other and to state-of-the-art methods on a collection of public data sets with up to 100 variables.

Bülent Yener Department of Computer Science Rensselaer Polytechnic Institute 110 8th St, Troy, NY, USA yener@cs.rpi.edu Pre-trained contextual language models are ubiquitously employed for language understanding tasks, but are unsuitable for resource-constrained systems. Noncontextual word embeddings are an efficient alternative in these settings. Such methods typically use one vector to encode multiple different meanings of a word, and incur errors due to polysemy. This paper proposes a two-stage method to distill multiple word senses from a pre-trained language model (BERT) by using attention over the senses of a word in a context and transferring this sense information to fit multi-sense embeddings in a skip-gram-like framework. We demonstrate an effective approach to training the sense disambiguation mechanism in our model with a distribution over word senses extracted from the output layer embeddings of BERT. Experiments on the contextual word similarity and sense induction tasks show that this method is superior to or competitive with state-of-the-art multi-sense embeddings on multiple benchmark data sets, and experiments with an embedding-based topic model (ETM) demonstrates the benefits of using this multi-sense embedding in a downstream application. While modern deep contextual word embeddings have dramatically improved the state-of-the-art in natural language understanding (NLU) tasks, shallow noncontextual representation of words are more practical solution in settings constrained by compute power or latency. In single-sense embeddings such as word2vec or GloVe, the different meanings of a word are represented by the same vector, which leads to the meaning conflation problem in the presence of polysemy.

Often the best way to learn a language is to immerse yourself in an environment where people speak it. The constant exposure, along with the pressure to communicate, helps you swiftly pick up and practice new vocabulary. But not everyone gets the opportunity to live or study abroad. In a new collaboration with IBM Research, Rensselaer Polytechnic Institute (RPI), a university based in Troy, New York, now offers its students studying Chinese another option: a 360-degree virtual environment that teleports them to the busy streets of Beijing or a crowded Chinese restaurant. Students get to haggle with street vendors or order food, and the environment is equipped with different AI capabilities to respond to them in real time.

An artificial intelligence (AI) algorithm can transform low-dose CT (LDCT) scans into high-quality exams that radiologists may even prefer over LDCT studies produced via commercial iterative reconstruction techniques (Nature Machine Intelligence 10.1038/s42256-019-0057-9). A team of researchers from Rensselaer Polytechnic Institute (RPI) in Troy, NY, and Massachusetts General Hospital (MGH) in Boston developed a deep-learning model called a modularized adaptive processing neural network (MAP-NN), which progressively reduces noise on LDCT images with guidance from the radiologist until the optimal level of image quality is achieved. Testing on images from three different vendors, three radiologists found the algorithm produced images that were either better or comparable to images processed with iterative reconstruction. The deep-learning method also processed images much faster. "The deep-learning approach can thus already effectively compete with iterative reconstruction solutions and potentially replace the iterative reconstruction approach," wrote the group, led by Hongming Shan of RPI.

As large eye-tracking datasets are created, data privacy is a pressing concern for the eye-tracking community. De-identifying data does not guarantee privacy because multiple datasets can be linked for inferences. A common belief is that aggregating individuals' data into composite representations such as heatmaps protects the individual. However, we analytically examine the privacy of (noise-free) heatmaps and show that they do not guarantee privacy. We further propose two noise mechanisms that guarantee privacy and analyze their privacy-utility tradeoff. Analysis reveals that our Gaussian noise mechanism is an elegant solution to preserve privacy for heatmaps. Our results have implications for interdisciplinary research to create differentially private mechanisms for eye tracking.

Data science relies on pipelines that are organized in the form of interdependent computational steps. Each step consists of various candidate algorithms that maybe used for performing a particular function. Each algorithm consists of several hyperparameters. Algorithms and hyperparameters must be optimized as a whole to produce the best performance. Typical machine learning pipelines consist of complex algorithms in each of the steps. Not only is the selection process combinatorial, but it is also important to interpret and understand the pipelines. We propose a method to quantify the importance of different components in the pipeline, by computing an error contribution relative to an agnostic choice of computational steps, algorithms and hyperparameters. We also propose a methodology to quantify the propagation of error from individual components of the pipeline with the help of a naive set of benchmark algorithms not involved in the pipeline. We demonstrate our methodology on image classification pipelines. The agnostic and naive methodologies quantify the error contribution and propagation respectively from the computational steps, algorithms and hyperparameters in the image classification pipeline. We show that algorithm selection and hyperparameter optimization methods like grid search, random search and Bayesian optimization can be used to quantify the error contribution and propagation, and that random search is able to quantify them more accurately than Bayesian optimization. This methodology can be used by domain experts to understand machine learning and data analysis pipelines in terms of their individual components, which can help in prioritizing different components of the pipeline.

Apple has agreed to pay $24.9 million to a patent holding company to resolve a 5-year-old lawsuit accusing Siri of infringing one of its patents. Apple will pay the money to Marathon Patent Group, the parent company of Texas firm Dynamic Advances, which held an exclusive license to a 2007 patent covering natural language user interfaces for enterprise databases. Marathon reported the settlement in a filing with the U.S. Securities and Exchange Commission Tuesday. On Wednesday, in response to the settlement, Magistrate Judge David Peebles of U.S. District Court for the Northern District of New York dismissed a lawsuit against Apple filed by Dynamic Advances and Rensselaer Polytechnic Institute in Troy, New York, where the natural language technology was created. A trial had been scheduled to begin early next month in Syracuse, New York.