Investigates the limits of cost-sensitive classification problems.
problem Cost-sensitive classification problem in mission-critical applications.
method Extending the minimax lower bound of balanced binary classification problem.
result Cost terms significantly impact the hardness of the problem.
New measure assesses deep neural networks' robustness to adversarial attacks.
problem Deep learning's fragility to adversarial attacks limits its adoption in mission-critical applications.
method Introduces residual error as a new performance measure for assessing adversarial robustness.
result Demonstrates effectiveness of residual error in assessing robustness of deep neural networks.
Survey on understanding neural networks for medical applications.
problem Black-box nature of deep neural networks hinders their use in critical applications.
method Comprehensive review of interpretability studies in neural networks.
result Interpretability research is crucial for the acceptance of neural networks in medical diagnosis.
Proposes a Bayesian approach to explain, justify, and quantify uncertainty in DNNs.
problem Lack of transparency and confidence in DNNs for critical applications.
method Bayesian approach to extract explanations, justifications, and uncertainty estimates from black box DNNs.
result Improves interpretability and reliability of DNNs, validated on CIFAR-10.
Knowledge distillation simplifies deep models into interpretable decision trees.
problem Interpretability of deep neural networks is challenging and important for practical deployment.
method Knowledge distillation applied to transform deep models into decision trees.
result The student model achieves better accuracy than vanilla decision trees.
Paper proposes a method to estimate individual treatment effects reliably from observational data.
problem Estimating individual treatment effects from observational data is challenging and important.
method The approach uses the Information Bottleneck principle to find more reliable representations for ITE estimation.
result The proposed model achieves state-of-the-art results and provides more reliable prediction performances with uncertainty information.
New algorithm robustly solves blind deconvolution problems.
problem Robustly solving blind deconvolution problems in the presence of noise and perturbations.
method Mirror Descent algorithm for robust continuous optimization.
result Provable robustness and convergence guarantees for the algorithm.
Deep RL predicts equipment maintenance from sensor data.
problem Equipment downtime due to sensor data overload.
method Model-free Deep Reinforcement Learning for optimal maintenance policy.
result Automatic maintenance policy learning from sensor data.
This paper simplifies deep ReLU networks into local linear models for better interpretability.
problem Limited transparency and interpretability of deep neural networks, especially ReLU networks.
method Local linear representation and equivalent set of local linear models (LLMs).
result Simplified deep ReLU networks for better interpretability and diagnostics.
Robots adapt to damage with a single policy and diagnosis.
problem Robotic failure due to damage during mission-critical tasks.
method Damage-aware control architecture using supervised learning for diagnosis and policy adaptation.
result Single-shot diagnosis and adaptation achieved with a single policy.
SPINN optimizes neural network inference on devices and cloud.
problem Inference on mobile devices is challenging due to high computational demands and dynamic connectivity.
method Synergistic progressive inference with a novel scheduler.
result SPINN achieves up to 2x higher throughput and reduces server cost by up to 6.8x.
New method for efficient uncertainty quantification in DeepONets.
problem Efficient uncertainty quantification for DeepONets with limited and noisy data.
method Ensemble Kalman Inversion (EKI) for ensembles of DeepONets.
result Improved uncertainty estimates for DeepONet predictions.
VINE visualizes statistical interactions in complex models.
problem Lack of utilities for regional explanations in black box models.
method VINE algorithm to extract and visualize statistical interaction effects.
result VINE provides a novel evaluation metric for visualizations.
Paper presents attacks on real-time object detection systems.
problem Adversarial attacks on real-time object detection systems.
method Three targeted adversarial Objectness Gradient attacks (TOG).
result Adversarial attacks can cause object-vanishing, object-fabrication, and object-mislabeling.
This paper tackles URLLC in 6G networks with deep learning.
problem Stringent requirements on end-to-end delay and reliability for mission-critical applications.
method Develops a multi-level architecture combining theoretical models and real-world data, using deep transfer learning and federated learning.
result Demonstrates improved performance in URLLC for mission-critical applications.
Generative adversarial networks enable distributed IoT IDS without central controller.
problem Detecting cyber attacks in IoT systems with privacy preservation.
method Proposes a fully distributed GAN-based IDS for IoT.
result Higher accuracy and lower false positive rate compared to standalone IDS.
This paper characterizes adversarial examples in deep learning.
problem Security threats posed by adversarial attacks in deep learning systems.
method Statistical characterization of adversarial examples, easy and hard categorization of attacks, extensive experimental study.
result Adversarial attacks behave differently under different hyperparameters and frameworks.
CoRMF uses RNNs to solve Ising models efficiently by ordering critical edges.
problem Solving Ising models efficiently and accurately.
method Criticality-ordered spin sequence and RNNs for mean-field factorization.
result Proves tighter error bounds than naive mean-field.
This paper tackles collision avoidance for many UAVs using MFG and ML.
problem Collision avoidance for many UAVs in real-time missions.
method Mean-field game (MFG) theory combined with machine learning (ML) to reduce computation and communication energy.
result The proposed MFG learning control method achieves collision avoidance with low communication and acceptable computation energy.
Guidelines for using explainable ML to avoid misuse.
problem Misuse of explainable ML, especially for harmful purposes.
method Proposed guidelines to promote best practices.
result Promote interpretable models and testing methods.
This work makes deep neural networks more resilient to adversarial attacks.
problem Making deep neural networks more resilient to adversarial attacks in real-time systems.
method Used GTSRB dataset to craft adversarial samples, then attacked a Deep Convolutional Neural Network to build a more resilient network.
result Built a more robust deep neural network that can resist adversarial attacks.
Few-shot learning improves bearing fault diagnosis with limited data.
problem Challenges in collecting sufficient fault data for robust classifier training.
method Model-Agnostic Meta-Learning (MAML) for few-shot learning.
result Framework achieves up to 25% higher accuracy than Siamese network.
Last-layer approximation improves UQ performance without sacrificing computational efficiency.
problem Epistemic uncertainty quantification for deep neural networks.
method Comparison of full-network and last-layer linearization using theoretical and empirical approaches.
result Last-layer approximation yields comparable UQ performance with improved computational efficiency.
This paper assesses neural network robustness under neuron failures.
problem Understanding neural network limits and capabilities in critical applications.
method Proved an upper bound on expected error due to neuron crashes, involving network parameters and Lipschitz coefficients.
result Estimates robustness without testing all failure configurations or accessing training data.
Bayesian approach predicts molecule bioactivity with uncertainty, aiding active learning.
problem Predicting bioactivity of small molecules accurately and reliably.
method Bayesian semi-supervised graph convolutional neural networks.
result Bayesian approach estimates uncertainty and enables active learning.
Research evaluates data poisoning attacks on regression learning and introduces a new defense strategy.
problem Data poisoning attacks on regression learning threaten model integrity in critical systems.
method Realistic scenarios, novel black-box attack, and evaluation on 26 datasets.
result Mean squared error (MSE) increases to 150% with only 2% poisoned samples.
This paper proposes using neural network sentiment measures to prioritize test inputs.
problem High cost of obtaining test oracle data limits testing efficacy of neural networks.
method Proposes using confidence, uncertainty, and surprise measures to prioritize test inputs.
result Sentiment measures can effectively flag inputs revealing DNN weaknesses.
Higgs bundles used in new applications.
problem None explicitly stated in the abstract.
method Overview of recent applications.
result Applications of Higgs bundles.
Android and Facebook provide third-party applications with access to users' private data and the ability to perform potentially sensitive operations (e.g., post to a user's wall or place phone calls). As a security measure, these platforms restrict applications' privileges with permission systems: users must approve th…
Integrates deep learning with existing apps using an estimator.
problem Lack of data for deep learning models to learn from existing applications.
method Estimate and Replace method: embeds estimator as DNN into base network, replaces at inference.
result Trained DNN with less data and outperformed non-interacting DNN.
New methods improve autotuning of exascale applications by 1.5x.
problem Finding optimal performance parameters for exascale applications.
method Multitask and transfer learning for autotuning.
result Average 1.5x improvement in application runtime.
DeepPlace learns to place applications in clusters using RL.
problem Manual placement rules for scheduling are non-trivial and suboptimal.
method Uses Deep Reinforcement Learning to learn optimal placement rules.
result Reduces resource competition and optimizes cluster utilization.
Curved flats linked to pairs of Lie applicable surfaces.
problem Understanding curved flats in Lie sphere geometry.
method One-to-one correspondence with pairs of Demoulin families of Lie applicable surfaces via Darboux transformation.
result Curved flats correspond to specific Lie applicable surface pairs.
Interactive applications incorporating high-data rate sensing and computer vision are becoming possible due to novel runtime systems and the use of parallel computation resources. To allow interactive use, such applications require careful tuning of multiple application parameters to meet required fidelity and latency …
CactusNets measure how useful features are for specific classes.
problem Lack of a metric for measuring how applicable learned features are to specific classes.
method Propose a metric for feature applicability and use it to estimate input applicability.
result Developed a new method for unsupervised learning called CactusNet.
An overview of some of the recent developments in the theory of valuations on convex sets and its generalizations to manifolds is given. The exposition is focused towards applications to integral geometry; several of such applications are discussed.
The aim of this paper is to discuss some applications of general topology in computer algorithms including modeling and simulation, and also in computer graphics and image processing. While the progress in these areas heavily depends on advances in computing hardware, the major intellectual achievements are the algorit…
Survey of deep RL in intelligent transportation systems.
problem Optimizing traffic signals and autonomous driving using deep RL.
method Comprehensive review of deep RL applications in traffic control and autonomous driving.
result Summarizes existing works in deep RL-based transportation applications.
Adjustment reduces bias in widely applicable Bayesian information criterion.
problem Overestimation of widely applicable Bayesian information criterion.
method Identified and adjusted an overestimating term in the criterion.
result Asymptotically unbiased estimator of log marginal likelihood.
The study proves estimates for transverse nonlinear equations on Sasakian manifolds with applications in geometry.
problem Estimating transverse fully nonlinear equations on Sasakian manifolds.
method Proving a priori estimates for transverse fully nonlinear equations.
result The study proves estimates for transverse fully nonlinear equations on Sasakian manifolds and gives geometric applications.
TUV Austria proposes certification for ML applications to ensure reliability.
problem Ensuring trust in AI applications to meet societal reliance requirements.
method Holistic approach analyzing security, functionality, data quality, ethics, and criticality levels.
result Certification process for low-risk ML applications in supervised learning.
The study characterizes polynomial conserved quantities for Lie applicable surfaces.
problem Characterizing polynomial conserved quantities for Lie applicable surfaces.
method Gauge theoretic approach for Lie applicable surfaces, including isothermic, Guichard, and L-isothermic surfaces. result Induced transformations of Lie applicable surfaces for well-known transformations and new Bäcklund-type transformation for linear Weingarten surfaces.
MetaDVFS uses device and application metadata to improve DVFS efficiency.
problem Improving energy efficiency in mobile platforms with diverse applications and hardware.
method Formulates DVFS as a multi-task reinforcement learning problem and introduces MetaDVFS, leveraging metadata for knowledge transfer.
result MetaDVFS achieves up to 26% improvement in Quality of Experience and up to 17% improvement in Performance-Power Ratio.
Ray is a distributed system for AI applications that learn from continuous interactions.
problem Demanding systems requirements for next-gen AI applications.
method Unified interface for task-parallel and actor-based computations, distributed scheduler, fault-tolerant store.
result Demonstrated scaling beyond 1.8 million tasks per second and better performance for reinforcement learning.
Isoparametric hypersurfaces and their application to special geometries
This paper reviews Douglas curvature in Finsler geometry.
problem Exploring Douglas curvature in Finsler spaces.
method Historical review, characterizations, generalizations, and applications.
result Significance and applications of Douglas curvature in Finsler geometry.
A New Trinomial Recombination Tree Algorithm and Its Applications
Survey of statistical queries and their applications.
problem Understanding statistical queries and their applications.
method Exploration of statistical queries model, definitions, and connections to learnability.
result Connections to learnability and applications in optimization, evolvability, and differential privacy.