Study improves RF sensor robustness for target recognition.
problem Variability in RF target responses makes them vulnerable to attacks.
method Evaluates techniques for building robust classification architectures.
result Improves accuracy in identifying true target characteristics.
UDA learns target domain from unlabeled data via source knowledge transfer.
problem Learning unlabeled target domain data.
method Transferring both source knowledge and target-relatedness.
result Improves UDA performance by simultaneous knowledge transfer.
Bayesian fusion improves radar target recognition for UAVs.
problem Improving radar target recognition for UAVs using multistatic radar configurations.
method Proposes a fully Bayesian RATR framework using Optimal Bayesian Fusion (OBF) to aggregate classification probability vectors from multiple radars.
result Empirical results show that the OBF method significantly enhances classification accuracy compared to other fusion methods and single radar configurations.
VoiceFilter separates target speaker from multi-speaker signals.
problem Speech recognition in multi-speaker environments.
method Speaker recognition network and spectrogram masking network trained together.
result Significant reduction in speech recognition WER on multi-speaker signals.
This paper creates imperceptible, effective audio adversarial examples for speech recognition.
problem Current adversarial examples in speech recognition are easily detectable and ineffective in real-world settings.
method Developed imperceptible audio adversarial examples using psychoacoustic masking and simulated environmental distortions.
result Successfully created imperceptible, targeted adversarial examples for speech recognition that are effective in real-world settings.
Personal VAD detects target speaker voice activity efficiently.
problem Efficiently detect target speaker voice activity for reduced computational cost and battery usage.
method Trains a neural network conditioned on speaker embedding or verification score, outputs probabilities for three speech classes.
result Trained model with 130K parameters outperforms combined standard VAD and speaker recognition networks.
Develops a deep generative model for radar target recognition using HRRP data.
problem Automatic target recognition in radar systems using high-resolution range profiles.
method Recurrent gamma belief network (rGBN) with hybrid stochastic-gradient MCMC and variational inference.
result Efficient and accurate classification with interpretable latent structure.
Real-time personalization for HAR models learns from new users without prior data.
problem Poor performance of HAR models on new users without labeled data.
method Incremental online domain adaptation using batch normalization.
result Personalized HAR models adapt to new users in real-time.
VoiceFilter-Lite separates speech from background in real-time for on-device speech recognition.
problem Separate speech from background in real-time for on-device speech recognition.
method Asymmetric loss, adaptive runtime suppression, quantization to 8-bit.
result VoiceFilter-Lite achieves real-time speech separation and maintains speech recognition performance.
This work tackles domain shift in speech emotion recognition by proposing class-wise adversarial domain adaptation.
problem Domain shift between corpora poses a challenge for speech emotion recognition, especially for positive/negative emotions.
method Class-wise adversarial domain adaptation to reduce shift between different corpora.
result Our method is effective even with limited target labeled examples, as demonstrated on EMODB and Aibo corpora.
Study shows mimicry attacks fail even with ASV-assisted target selection.
problem Can mimicry attacks be successful with ASV-assisted target selection?
method Used ASV to select target speakers for mimicry attacks and tested with x-vector system.
result Mimics did not succeed in spoofing the x-vector system, but relative ordering of targets was consistent.
Researchers create adversarial examples to deceive iris recognition systems.
problem Tackling the vulnerability of iris recognition systems to adversarial attacks.
method Developed a deep auto-encoder surrogate network to generate adversarial examples for iris recognition systems.
result Demonstrated that adversarial examples can fool iris recognition systems in both white-box and black-box settings.
Morph accelerates 3D CNNs for video recognition, reducing energy consumption and improving performance.
problem Efficiently accelerating 3D CNNs for video recognition is challenging due to their large memory footprint and higher dimensionality.
method Designing a flexible accelerator called Morph that adapts to different spatial and temporal tiling strategies, and codesigning a software infrastructure to control the hardware.
result Morph achieves up to 3.4x reduction in energy consumption and up to 5.1x improvement in performance/watt compared to a baseline 3D CNN accelerator.
The paper addresses bias in visual recognition models by reweighting observations.
problem Bias in deep neural networks trained on biased image databases.
method Reweighting observations based on known biasing mechanisms to form a nearly debiased estimator.
result The approach can remedy representativeness issues in visual recognition systems.
Speech Translation has always been about giving source text or audio input and waiting for system to give translated output in desired form. In this paper, we present the Acoustic Dialect Decoder (ADD) - a voice to voice ear-piece translation device. We introduce and survey the recent advances made in the field of Spee…
Deep architecture learns transferable features for robust speech emotion recognition.
problem Robust and discriminative features for diverse speech emotion domains.
method Jointly uses CNN for domain-shared features and LSTM for domain-specific emotion classification.
result Transferable features provide gains up to 18.4% in speech emotion recognition.
Proposes a new learning framework for scalable image classification.
problem Scalability issues in CNN-based image classifiers for large number of classes.
method Combines hierarchical classification and auxiliary learning to address scalability issues.
result Reduces classification errors by up to 3.56% on CIFAR-10 dataset.
DUPLE tackles cross-deployment recognition in fiber-optic perimeter security with meta-learning.
problem Cross-deployment recognition challenges in fiber-optic perimeter security due to label scarcity and distribution shifts.
method DUPLE employs statistically guided meta-learning to enhance recognition robustness across unseen deployments.
result DUPLE consistently outperforms traditional and meta-learning baselines in cross-deployment DFOS benchmarks.
Paper presents a method to train NER models without labelled data using weak supervision.
problem Dealing with NER performance drop in new domains without labelled data.
method Weak supervision through automatic annotation and hidden Markov model integration.
result Improvement of about 7 percentage points in entity-level F1 scores. Paper detects adversarial speech inputs with high accuracy.
problem Adversarial attacks on ASR systems.
method Uncertainty quantification using neural networks.
result Detection accuracy of adversarial inputs over 0.99.
V-BAD is a first black-box video attack framework that successfully fools deep video recognition models.
problem Vulnerability of video recognition models to black-box adversarial attacks.
method Tentative perturbations transferred from image models and partition-based rectifications for patches of tentative perturbations.
result V-BAD can craft both untargeted and targeted attacks with high success rates using a manageable number of queries.
Human activity recognition plays an important role in people's daily life. However, it is often expensive and time-consuming to acquire sufficient labeled activity data. To solve this problem, transfer learning leverages the labeled samples from the source domain to annotate the target domain which has few or none labe…
Robust ASR model removes fast-changing features to resist attacks.
problem Vulnerability of ASR systems to adversarial attacks.
method Removing fast-changing features using slow feature analysis or low-pass filtering.
result Hybrid ASR models are more than four times more robust against targeted attacks.
Supervised learning with large scale labeled datasets and deep layered models has made a paradigm shift in diverse areas in learning and recognition. However, this approach still suffers generalization issues under the presence of a domain shift between the training and the test data distribution. In this regard, unsup…
Transfer learning does not improve character recognition performance.
problem Improving character recognition performance using transfer learning.
method Performed experiments with varying levels of similarity between source and target tasks, transferring both parameters and features.
result No significant advantage gained by transfer learning over traditional machine learning.
Adversarial training improves audio classifier robustness.
problem Improving robustness of audio classifiers against adversarial attacks.
method Adversarial training of deep neural networks on audio data.
result Adversarially trained models can resist higher perturbations and fooling rates.
We present the recent advances along with an error analysis of the IBM speaker recognition system for conversational speech. Some of the key advancements that contribute to our system include: a nearest-neighbor discriminant analysis (NDA) approach (as opposed to LDA) for intersession variability compensation in the i-…
A new method trains deep neural networks for open set domain adaptation without negative open set difference.
problem Training deep neural networks for open set domain adaptation without negative open set difference.
method Proposes a new upper bound of target-domain risk, including source-domain risk, ε-open set difference (Δε), distributional discrepancy, and constant. Uses gradient descent for source-domain risk and Δε, and adversarial training for distributional discrepancy. Trains DNNs via minimizing the new upper bound. result Shows state-of-the-art performance on benchmark datasets.
Framework tracks and predicts multiple objects in autonomous driving.
problem Challenges in multi-target tracking due to object number fluctuation and occlusion.
method Constrained mixture sequential Monte Carlo (CMSMC) method with a mixture representation.
result Framework can track and predict multiple objects simultaneously without explicit data association.
In this paper, we propose a novel unsupervised domain adaptation algorithm based on deep learning for visual object recognition. Specifically, we design a new model called Deep Reconstruction-Classification Network (DRCN), which jointly learns a shared encoding representation for two tasks: i) supervised classification…
Paper shows attacks on transfer learning models without target-specific info.
problem Security vulnerability in transfer learning models.
method Brute force attack on pre-trained models without target-specific data.
result Demonstrates effective and efficient attacks on transfer learning models.
ADDoG improves cross-dataset speech emotion recognition.
problem Cross-dataset speech emotion recognition failure.
method ADDoG uses an iterative approach to move representations closer together across datasets.
result ADDoG and MADDoG improve cross-dataset speech emotion recognition.
SA-GAN improves HAR model performance across new users.
problem Poor performance of HAR models on new user data.
method Generative Adversarial Network (GAN) for cross-subject transfer learning.
result SA-GAN outperformed other methods in HAR tasks.
Proposes a new method to improve target annotation in ATR.
problem Challenges in annotating automatic target recognition due to lack of labeled data.
method Hybrid contrastive learning and cycle-consistency-based transductive transfer learning (C3TTL) framework.
result Significantly lower Fréchet Inception Distance (FID) score and improved performance in annotating civilian and military vehicles, as well as ship targets.
Graph autoencoders enable ML across diverse sensor networks.
problem Deploying ML across different sensor networks with varying types or layouts.
method Graph Autoencoders for activity recognition across heterogeneous sensor networks.
result Transferable activity classifiers achieve 75% accuracy on unseen sensor layouts.
Inspired by recent work on neural network image generation which rely on backpropagation towards the network inputs, we present a proof-of-concept system for speech texture synthesis and voice conversion based on two mechanisms: approximate inversion of the representation learned by a speech recognition neural network,…
Dynamic Transfer Networks improve NER from clinical notes with less data.
problem Improving named entity recognition from clinical notes with limited data.
method Dynamic Transfer Networks (DTN) using parameter sharing and gating architecture.
result DTN achieves better NER performance with fewer training settings.
Few-shot NER classification with limited data.
problem Limited annotated data for Named Entity Recognition.
method Prototypical Network + Transfer Learning
result Well-performing classifiers trained on only 20 instances.
This paper presents our approach to the One-Minute Gradual-Emotion Recognition (OMG-Emotion) Challenge, focusing on dimensional emotion recognition through visual analysis of the provided emotion videos. The approach is based on a Convolutional and Recurrent (CNN-RNN) deep neural architecture we have developed for the …
RAN model recognizes multiple activities from unlabeled sensor data.
problem Handling weakly labeled multi-activity data from wearable sensors.
method Recurrent Attention Networks (RAN) for sequential multi-activity recognition and localization.
result RAN model can infer multiple activities and determine activity locations from unlabeled data.
New method generates unrestricted adversarial face images to bypass robust face recognition systems.
problem Vulnerability of face recognition systems to unrestricted adversarial examples.
method Image translation techniques to generate large perturbations in face images.
result Achieved high attack success rates (90% and 80%) under white- and black-box settings. New approach predicts under latent shifts using high-dimensional images.
problem Prediction under latent subgroup shifts with high-dimensional observations.
method Recognition-parametrised model (RPM) for identifying causal latent structure.
result Successfully adapts predictions for high-dimensional image data.
Data-target pairing is an important step towards multi-target localization for the intelligent operation of unmanned systems. Target localization plays a crucial role in numerous applications, such as search, and rescue missions, traffic management and surveillance. The objective of this paper is to present an innovati…
Vehicle recognition and classification have broad applications, ranging from traffic flow management to military target identification. We demonstrate an unsupervised method for automated identification of moving vehicles from roadside audio sensors. Using a short-time Fourier transform to decompose audio signals, we t…
Despite the remarkable progress in face recognition related technologies, reliably recognizing faces across ages still remains a big challenge. The appearance of a human face changes substantially over time, resulting in significant intra-class variations. As opposed to current techniques for age-invariant face recogni…
Defines a similarity measure for classification distributions.
problem Measuring similarity between classification distributions.
method Proposes task similarity, a novel measure quantifying performance of source distributions on target distributions.
result Empirical task similarity correlates with transfer efficiency and semantic similarity of source distributions.
DARN uses multiple source datasets to adapt to a new target dataset.
problem Learning a model for a new, related dataset using multiple source datasets.
method DARN applies domain discrepancy minimization with a theoretical generalization bound to adjust source domain weights.
result DARN significantly outperforms state-of-the-art alternatives on real-world datasets.
The application of deep recurrent networks to audio transcription has led to impressive gains in automatic speech recognition (ASR) systems. Many have demonstrated that small adversarial perturbations can fool deep neural networks into incorrectly predicting a specified target with high confidence. Current work on fool…