Input-output Pairs Research Articles

Dialect Bias in Automatic Speech Recognition: Analysis of Appalachian English

The rise of artificial intelligence has prompted increased scrutiny of systemic biases in automatic speech recognition technologies. One focal topic of discussion has been the degraded performance for speakers of African American and Southern U.S. English. This study aims to contribute to the research on bias in voice-AI by investigating speech recognition performance for Appalachian English, an often-stigmatized variety in American society. Participants were recruited from Southern Appalachia (Eastern Tennessee), with a non-Southern Appalachian (Central Pennsylvania) sample included as a reference group. The participants read aloud the Goldilocks fairytale and the Rainbow Passage, and the recordings were processed using Dartmouth Linguistic Automation (DARLA). We conducted two sets of analyses on the vowel phonemes. The first analysis assessed DARLA’s effectiveness in recognizing vowels. The system returned higher phoneme error rates for Southern Appalachian speech compared to the non-Southern dataset. Next, we conducted a detailed error analysis on the misrecognized input-output phoneme pairs. The results suggested dialect bias in the system, with 50.2% of the errors in the Southern dataset attributed to participation in the Southern Vowel Shift. These findings underscore the importance of integrating sociolectal variation into the acoustic model to mitigate dialect bias for currently underserved users.

Enhancing performance of 3D point completion network using consistency loss

Point cloud completion networks are conventionally trained to minimize the disparities between the completed point cloud and the ground-truth counterpart. However, an incomplete object-level point cloud can have multiple valid completion solutions when it is examined in isolation. This one-to-many mapping issue can cause contradictory supervision signals to the network because the loss function may produce different values for identical input–output pairs of the network. In many cases, this issue could adversely affect the network optimization process. In this work, we propose to enhance the conventional learning objective using a novel completion consistency loss to mitigate the one-to-many mapping problem. Specifically, the proposed consistency loss ensure that a point cloud completion network generates a coherent completion solution for incomplete objects originating from the same source point cloud. Experimental results across multiple well-established datasets and benchmarks demonstrated the proposed completion consistency loss have excellent capability to enhance the completion performance of various existing networks without any modification to the design of the networks. The proposed consistency loss enhances the performance of the point completion network without affecting the inference speed, thereby increasing the accuracy of point cloud completion. Notably, a state-of-the-art point completion network trained with the proposed consistency loss can achieve state-of-the-art accuracy on the challenging new MVP dataset. The code and result of experiment various point completion models using proposed consistency loss will be available at: https://github.com/kaist-avelab/ConsistencyLoss .

Algorithm XXXX: The Delaunay Density Diagnostic

Accurate approximation of a real-valued function depends on two aspects of the available data: the density of inputs within the domain of interest and the variation of the outputs over that domain. There are few methods for assessing whether the density of inputs is sufficient to identify the relevant variations in outputs—i.e., the “geometric scale” of the function—despite the fact that sampling density is closely tied to the success or failure of an approximation method. In this paper, we introduce a general purpose, computational approach to detecting the geometric scale of real-valued functions over a fixed domain using a deterministic interpolation technique from computational geometry. The algorithm is intended to work on scalar data in moderate dimensions (2-10). Our algorithm is based on the observation that a sequence of piecewise linear interpolants will converge to a continuous function at a quadratic rate (in \(L^{2}\) norm) if and only if the data are sampled densely enough to distinguish the feature from noise (assuming sufficiently regular sampling). We present numerical experiments demonstrating how our method can identify feature scale, estimate uncertainty in feature scale, and assess the sampling density for fixed (i.e. static) datasets of input–output pairs. We include analytical results in support of our numerical findings and have released lightweight code that can be adapted for use in a variety of data science settings.

Security of End-to-End medical images encryption system using trained deep learning encryption and decryption network

The Internet of Medical Things (IoMT) links medical devices and wearable, enhancing healthcare. To secure sensitive patient data over the IoMT, encryption is vital to retain confidentiality, prevent tampering, ensure authenticity, and secure data transfer. The intricate neural network architecture of deep learning models adds a layer of complexity and non-linearity to the encryption process, rendering it highly resistant to plaintext attacks. Specifically, the Cycle_GAN network is used as the leading learning network. This work suggests deep learning-based encryption for medical images using Cycle_GAN, a Generative Adversarial Network. Cycle_GAN changes images without paired training data that improves quality and feature preservation. Unlike conventional image-to-image translation techniques, Cycle_GAN doesn’t require a dataset with corresponding input–output pairs. Traditional methods typically needs paired data to learn the mapping between input and output images. Paired data can be challenging to obtain, specifically in medical imaging where gathering and annotating data can be time-consuming, laborious and expensive. The use of Cycle_GAN overwhelms this constraint by using unpaired data, where the input and output images are not explicitly paired. This method ensures confidentiality, authenticity, and secure transfer. Cycle_GAN consists of two major components: a generator used to modify the images, and a discriminator used to distinguish between real and fake images. Further, the Binary-Cross Entropy loss function is employed to train the network for precise predictions. The experiments are carried out on skin cancer datasets. The results demonstrate high-level efficient, systematic and coherent encryption as compared with other modernized medical image encryption methods. The proposed technique offers several benefits, including efficient encryption and decryption and robustness against unauthorized access.

Multi-Fidelity Physics-Informed Generative Adversarial Network for Solving Partial Differential Equations

Abstract We propose a novel method for solving partial differential equations using multi-fidelity physics-informed generative adversarial networks. Our approach incorporates physics supervision into the adversarial optimization process to guide the learning of the generator and discriminator models. The generator has two components: one that approximates the low-fidelity response of the input and another that combines the input and low-fidelity response to generate an approximation of high-fidelity responses. The discriminator identifies whether the input–output pairs accord not only with the actual high-fidelity response distribution, but also with physics. The effectiveness of the proposed method is demonstrated through numerical examples and compared to existing methods.

On languaging a simulation engine: Rapid modeling of nanoporous media sorption by hierarchical language model

Language model intelligence is revolutionizing the way we program materials simulations. However, the diversity of simulation scenarios renders it challenging to precisely transform human language into a tailored simulator. Here, using three types of expert language model, we propose a language-to-simulation (Lang2Sim) framework that enables interactive navigation on languaging a simulation engine, by taking a scenario instance of water sorption in porous matrices. Unlike line-by-line coding of a target simulator, the language models interpret each simulator as an assembly of invariant tool function and its variant input–output pair. Lang2Sim enables the precise transform of textual description by functionalizing and sequentializing the language models of, respectively, rationalizing the tool categorization, customizing its input–output combinations, and distilling the simulator input into executable format. Importantly, depending on its expert type, each language model features a distinct processing of chat history to best balance its memory limit and information completeness, thus leveraging the model intelligence to unstructured nature of human request. Overall, this work establishes language model as an intelligent platform to unlock the era of languaging a simulation engine.

A Time-Consistency Curriculum for Learning From Instance-Dependent Noisy Labels.

Many machine learning algorithms are known to be fragile on simple instance-independent noisy labels. However, noisy labels in real-world data are more devastating since they are produced by more complicated mechanisms in an instance-dependent manner. In this paper, we target this practical challenge of Instance-Dependent Noisy Labels by jointly training (1) a model reversely engineering the noise generating mechanism, which produces an instance-dependent mapping between the clean label posterior and the observed noisy label and (2) a robust classifier that produces clean label posteriors. Compared to previous methods, the former model is novel and enables end-to-end learning of the latter directly from noisy labels. An extensive empirical study indicates that the time-consistency of data is critical to the success of training both models and motivates us to develop a curriculum selecting training data based on their dynamics on the two models' outputs over the course of training. We show that the curriculum-selected data provide both clean labels and high-quality input-output pairs for training the two models. Therefore, it leads to promising and robust classification performance even in notably challenging settings of instance-dependent noisy labels where many SoTA methods could easily fail. Extensive experimental comparisons and ablation studies further demonstrate the advantages and significance of the time-consistency curriculum in learning from instance-dependent noisy labels on multiple benchmark datasets.

Randomized compression of rank-structured matrices accelerated with graph coloring

A randomized algorithm for computing a data sparse representation of a given rank-structured matrix A (a.k.a. an H-matrix) is presented. The algorithm draws on the randomized singular value decomposition (RSVD), and operates under the assumption that methods for rapidly applying A and A∗ to vectors are available. The algorithm uses graph coloring algorithms to analyze the hierarchical tree that defines the rank structure to generate a tailored probability distribution from which to draw the random test matrices. The matrix is then applied to the test matrices, and in a final step the matrix itself is reconstructed by the observed input–output pairs. The method presented is an evolution of the “peeling algorithm” of Lin et al. (2011). For the case of uniform trees, the new method substantially reduces the pre-factor of the original peeling algorithm. More significantly, the new technique leads to dramatic acceleration for many non-uniform trees since it constructs test matrices that are optimized for a given tree. The algorithm is particularly effective for kernel matrices involving a set of points restricted to a lower dimensional object than the ambient space, such as a boundary integral equation defined on a surface in three dimensions.

Supervised Learning: An InDepth Analysis

Supervised learning pivotal machine learning paradigm wherein models are trained on labeled datasets. They predict outcomes or classify data. It includes methodologies and diverse applications of supervised learning. Emphasizing significance in modern technology. Key methodologies encompass linear regression logistic regression. Also decision trees, support vector machines neural networks. Each with unique advantages for specific tasks. Versatility is demonstrated through applications in image and speech recognition. Natural language processing, medical diagnosis and financial forecasting also highlighted. Challenges include overfitting data quality, interpretability. Scalability discussed emphasizing areas for ongoing research. Future directions point towards transfer learning. Integration with semisupervised and unsupervised methods. Explainable AI and quantum machine learning promise further enhancements. They will impact supervised learning across various domains. Key words: Supervised learning, Machine learning, Labeled datasets, Algorithm training, Outcome prediction, Data classification, Input-output pairs, Model generalization, Accurate predictions, Real-world applications, Spam filters, Recommendation systems, Medical diagnosis, Autonomous driving, Efficiency enhancement, Decision-making, User experience, Data handling.

Deep-Learning-Based Strong Ground Motion Signal Prediction in Real Time

Processing ground motion signals at early stages can be advantageous for issuing public warnings, deploying first-responder teams, and other time-sensitive measures. Multiple Deep Learning (DL) models are presented herein, which can predict triaxial ground motion accelerations upon processing the first-arriving 0.5 s of recorded acceleration measurements. Principal Component Analysis (PCA) and the K-means clustering algorithm were utilized to cluster 17,602 accelerograms into 3 clusters using their metadata. The accelerograms were divided into 1 million input–output pairs for training, 100,000 for validation, and 420,000 for testing. Several non-overlapping forecast horizons were explored (1, 10, 50, 100, and 200 points). Various architectures of Artificial Neural Networks (ANNs) were trained and tested, such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) networks, and CNN-LSTMs. The utilized training methodology applied different aspects of supervised and unsupervised learning. The LSTM model demonstrated superior performance in terms of short-term prediction. A prediction horizon of 10 timesteps in the future with a Root Mean Squared Error (RMSE) value of 8.43 × 10−6 g was achieved. In other words, the LSTM model exhibited a performance improvement of 95% compared to the baseline benchmark, i.e., ANN. It is worth noting that all the considered models exhibited acceptable real-time performance (0.01 s) when running in testing mode. The CNN model demonstrated the fastest computational performance among all models. It predicts ground accelerations under 0.5 ms on an Intel Core i9-10900X CPU (10 cores). The models allow for the implementation of real-time structural control responses via intelligent seismic protection systems (e.g., magneto-rheological (MR) dampers).

A Comprehensive Discourse on Shallow Learning and its Applications

Shallow learning, a fundamental approach in machine learning, encompasses a variety of algorithms and techniques aimed at learning patterns and making predictions from labelled data. Unlike deep learning, which involves complex architectures with multiple layers of abstraction, shallow learning focuses on simpler models with limited complexity. This abstract explores the essence of shallow learning, its algorithms, applications, and challenges. Shallow learning algorithms include classic methods such as decision trees, support vector machines, k-nearest neighbours, and logistic regression, among others. These algorithms are typically trained using supervised learning techniques, where the model learns from input-output pairs to make predictions on new, unseen data. Shallow learning models excel in tasks such as classification and regression, where the goal is to assign labels or predict continuous values to input data. Applications of shallow learning span across various domains, including healthcare, finance, marketing, and cyber security. In healthcare, shallow learning models are used for disease diagnosis and prognosis prediction based on patient data. In finance, these models aid in fraud detection, credit scoring, and stock market prediction. Marketing applications involve customer segmentation and churn prediction, while in cyber security; shallow learning is utilized for malware detection and network intrusion detection.

Generative deep learning for the development of a type 1 diabetes simulator

BackgroundType 1 diabetes (T1D) simulators, crucial for advancing diabetes treatments, often fall short of capturing the entire complexity of the glucose-insulin system due to the imprecise approximation of the physiological models. This study introduces a simulation approach employing a conditional deep generative model. The aim is to overcome the limitations of existing T1D simulators by synthesizing virtual patients that more accurately represent the entire glucose-insulin system physiology.MethodsOur methodology utilizes a sequence-to-sequence generative adversarial network to simulate virtual T1D patients causally. Causality is embedded in the model by introducing shifted input-output pairs during training, with a 90-min shift capturing the impact of input insulin and carbohydrates on blood glucose. To validate our approach, we train and evaluate the model using three distinct datasets, each consisting of 27, 12, and 10 T1D patients, respectively. In addition, we subject the trained model to further validation for closed-loop therapy, employing a state-of-the-art controller.ResultsThe generated patients display statistical similarity to real patients when evaluated on the time-in-range results for each of the standard blood glucose ranges in T1D management along with means and variability outcomes. When tested for causality, authentic causal links are identified between the insulin, carbohydrates, and blood glucose levels of the virtual patients. The trained generative model demonstrates behaviours that are closer to reality compared to conventional T1D simulators when subjected to closed-loop insulin therapy using a state-of-the-art controller.ConclusionsThese results highlight our approach’s capability to accurately capture physiological dynamics and establish genuine causal relationships, holding promise for enhancing the development and evaluation of therapies in diabetes.

A machine learning model for reconstructing skin-friction drag over ocean surface waves

In order to improve the predictive abilities of weather and climate models, it is essential to understand the behaviour of wind stress at the ocean surface. Wind stress is contingent on small-scale interfacial dynamics typically not directly resolved in numerical models. Although skin friction contributes considerably to the total stress up to moderate wind speeds, it is notoriously challenging to measure and predict using physics-based approaches. This work proposes a supervised machine learning (ML) model that estimates the spatial distribution of the skin-friction drag over wind waves using solely wave elevation and wave age, which are relatively easy to acquire. The input–output pairs are high-resolution wave profiles and their corresponding surface viscous stresses collected from laboratory experiments. The ML model is built upon a convolutional neural network architecture that incorporates the Mish nonlinearity as its activation function. Results show that the model can accurately predict the overall distribution of viscous stresses; it captures the peak of viscous stress at/near the crest and its dramatic drop to almost null just past the crest in cases of intermittent airflow separation. The predicted area-aggregate skin friction is also in excellent agreement with the corresponding measurements. The proposed method offers a practical pathway for estimating both local and area-aggregate skin friction and can be easily integrated into existing numerical models for the study of air–sea interactions.

Online machine learning-based channel estimation for underwater acoustic communications

In recent years, machine learning-assisted underwater acoustic (UWA) communication technology has been proved its promising performance in simulations and ideal training and testing environments. However, given the unique characteristics of real-world UWA communication, offline training-based methods face two major challenges: (1) The severe temporal and spatial variations result in a significant gap between offline training and actual deployment conditions. (2) The availability of high-quality labeled samples is extremely limited, making it difficult to construct an appropriate offline training sample set. To address this, we propose two online learning methods for UWA channel estimation and apply them to an OFDM communication system. The first method employs an unsupervised learning approach, where pilot signals are split and input-output pairs for model construction are formed. Thus, the received real-time data can be applied for the online training directly. The second method adjusts the loss function during the machine learning training phase to embed the learned channel within it. It allows online-sampled data to be trained sequentially by leveraging prior knowledge of transmitted information as the actual learning target. Based on the experimental results of measured underwater acoustic channels, it is demonstrated that the proposed online learning methods can approximate the near-optimal solution.

Incremental Stability of Discrete-Time First-Order Projection Elements

In this paper, we present and analyze a class of discrete-time and sampled-data controllers named first-order projection elements (FOPEs). FOPEs form a class of hybrid controllers that have an underlying linear base dynamics, which are combined with a projection operator that ensures that the input-output (I/O) pair is contained in a well-designed sector. The I/O sector-constraint enables performance benefits in various application domains including motion control of high-precision machines. After the introduction of FOPEs, we focus on incremental stability analysis of these hybrid controllers. We show that the properties of the projection operator ensure incremental stability of the open-loop FOPE dynamics in case of an underlying linear dynamics that is exponentially stable. In this case, we compute tight bounds on the incremental gain. Under an appropriate small-gain condition, we show incremental input-to-state stability of the sampled-data interconnection of a FOPE with a continuous-time linear time-invariant plant. For the case of projected integrator dynamics, we provide a counterexample showing that incremental stability can be compromised. We provide additional, necessary and sufficient conditions under which FOPEs with integrator dynamics are incrementally asymptotically stable. Numerical simulations provide a practical example of the established incremental properties for the trajectories of a dynamical system coming from an industrial motion control application.

Adaptive and Robust Direct Data-Driven Controller for Surface Vessel Navigation

The challenging and unpredictable nature of the marine environment poses challenges for both human operators and non adaptive automatic controllers. This paper introduces and applies an extension of the DeePC algorithm as the vessel controller, aligning with the increasing trend of data-driven reference tracking in control system theory. The algorithm records input-output pairs during vessel operation, constructs a linear model that mimics the recorded system's trajectory and uses filtered and corrected trajectory required for the future behavior prediction in the direct data-driven scheme. The resulting controller is closed-loop stable, outperforms conventional control methods and is able to adapt to changing model parameters over time. The method is evaluated on a simulated Unmanned Surface Vehicle (USV), where it is initially used as a velocity controller and then extended to a dynamic positioning controller.

A tutorial on the Bayesian statistical approach to inverse problems

Inverse problems are ubiquitous in science and engineering. Two categories of inverse problems concerning a physical system are (1) estimate parameters in a model of the system from observed input–output pairs and (2) given a model of the system, reconstruct the input to it that caused some observed output. Applied inverse problems are challenging because a solution may (i) not exist, (ii) not be unique, or (iii) be sensitive to measurement noise contaminating the data. Bayesian statistical inversion (BSI) is an approach to tackle ill-posed and/or ill-conditioned inverse problems. Advantageously, BSI provides a “solution” that (i) quantifies uncertainty by assigning a probability to each possible value of the unknown parameter/input and (ii) incorporates prior information and beliefs about the parameter/input. Herein, we provide a tutorial of BSI for inverse problems by way of illustrative examples dealing with heat transfer from ambient air to a cold lime fruit. First, we use BSI to infer a parameter in a dynamic model of the lime temperature from measurements of the lime temperature over time. Second, we use BSI to reconstruct the initial condition of the lime from a measurement of its temperature later in time. We demonstrate the incorporation of prior information, visualize the posterior distributions of the parameter/initial condition, and show posterior samples of lime temperature trajectories from the model. Our Tutorial aims to reach a wide range of scientists and engineers.

Width residual neuro fuzzy system based on random feature selection

To shorten the operating time of the high-dimensional problems on fuzzy systems, we proposed the width residual neuro fuzzy system (WRNFS) before, but the discussion on the structure of WRNFS was insufficient, especially on the divide-and-conquer strategies of the input dimensions. In previous research, the optimization methods for WRNFS were not discussed. In this paper we proposed the first optimization method for WRNFS, which is an improved scheme for grouping the input dimensions of WRNFS, using random feature selection(RFS) to find a better solution, so as to improve the overall capability of the system. We call the width residual neuro fuzzy system based on random feature-selection as RFS-WRNFS. In this paper, the exhaustive experiment analysis and practical test of WRNFS and RFS-WRNFS are carried out on the reconstructed MG dataset, and the following conclusions are obtained: ding172 The performance of WRNFS is generally consistent when the structure of the WRNFS sub-systems and the input-output pairs are fixed; ding173 When searching for the optimal solution on the WRNFS, the time cost of exhaustive search is acceptable when the system remains in a small scale; ding174 In most cases, RFS-WRNFS carries out several random tests and produces better results than WRNFS. Furthermore, assuming that the input dimension is N and the times of attempts used to random feature selection for a better solution of WRNFS is M, we found:1) when M = 1 N, there is a certain probability to get an acceptable solution, and the system takes the shortest time; 2) When M = 2 N, there is a great chance to get an acceptable solution in a limited time; 3) When M = 3 N, best solution can be obtained with the longest search time. We suggest M = 2 N for the RFS-WRNFS for the comprehensive performance. Comparing the experiment results of exhaustive search and random feature selection, WRNFS always reaches the optimal solution by exhaustive search through a finite set in a limited time, while RFS-WRNFS in most time keeps a good balance between prediction precision and time efficiency.

Complexity Analysis of the SAT Attack on Logic Locking

Due to the adoption of horizontal business models following the globalization of semiconductor manufacturing, the overproduction of integrated circuits (ICs) and the piracy of intellectual properties (IPs) can lead to significant damage to the integrity of the semiconductor supply chain. Logic locking emerges as a primary design-for-security measure to counter these threats, where ICs become fully functional only when unlocked with a secret key. However, Boolean satisfiability-based attacks have rendered most locking schemes ineffective. This gives rise to numerous defenses and new locking methods to achieve SAT resiliency. This paper provides a unique perspective on the SAT attack efficiency based on conjunctive normal form (CNF) stored in SAT solver. First, we show how the attack learns new relations between keys in every iteration using distinguishing input patterns and the corresponding oracle responses. The input-output pairs result in new CNF clauses of unknown keys to be appended to the SAT solver, which leads to an exponential reduction in incorrect key values. Second, we demonstrate that the SAT attack can break any locking scheme within linear iteration complexity of key size. Moreover, we show how key constraints on point functions affect the SAT attack complexity. We explain why proper key constraint on AntiSAT reduces the complexity effectively to constant 1. The same constraint helps the breaking of CAS-Lock down to linear iteration complexity. Our analysis provides a new perspective on the capabilities of SAT attack against multiplier benchmark c6288, and we provide new directions to achieve SAT resiliency.

Local look-up table upsampling for accelerating image processing

The resolution of cameras is increasing, and speedup of various image processing is required to accompany this increase. A simple way of acceleration is processing the image at low resolution and then upsampling the result. Moreover, when we can use an additional high-resolution image as guidance formation for upsampling, we can upsample the image processing results more accurately. We propose an approach to accelerate various image processing by downsampling and joint upsampling. This paper utilizes per-pixel look-up tables (LUTs), named local LUT, which are given a low-resolution input image and output pair. Subsequently, we upsample the local LUT. We can then generate a high-resolution image only by referring to its local LUT. In our experimental results, we evaluated the proposed method on several image processing filters and applications: iterative bilateral filtering, ℓ0\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\ell _0$$\\end{document} smoothing, local Laplacian filtering, inpainting, and haze removing. The proposed method accelerates image processing with sufficient approximation accuracy, and the proposed outperforms the conventional approaches in the trade-off between accuracy and efficiency. Our code is available at https://fukushimalab.github.io/LLF/.