arXiv research

We propose a novel approach to addressing the vanishing (or exploding) gradient problem in deep neural networks. We construct a new architecture for deep neural networks where all layers (except the output layer) of the network are a combination of rotation, permutation, diagonal, and activation sublayers which are all…

Active learning holds promise of significantly reducing data annotation costs while maintaining reasonable model performance. However, it requires sending data to annotators for labeling. This presents a possible privacy leak when the training set includes sensitive user data. In this paper, we describe an approach for…

Tempered sigmoids improve deep learning privacy.

IMPACT optimizes LLM compression by focusing on activation importance, reducing model size up to 55.4%.

This paper converts NACE classification into embeddings to preserve hierarchical structure.

Binary autoencoder with sparse hidden layer preserves information and zero reconstruction error.

State-of-the-art subspace clustering methods are based on expressing each data point as a linear combination of other data points while regularizing the matrix of coefficients with $\ell_1$, $\ell_2$ or nuclear norms. $\ell_1$ regularization is guaranteed to give a subspace-preserving affinity (i.e., there are no conne…

New initialization schemes preserve fractional moments of weights in deep networks, improving training and test performance.

Traditionally, the automatic recognition of human activities is performed with supervised learning algorithms on limited sets of specific activities. This work proposes to recognize recurrent activity patterns, called routines, instead of precisely defined activities. The modeling of routines is defined as a metric lea…

Improved diffusion models for image synthesis with better training dynamics.

Active learning improves GP regression on complex, high-dimensional data.

Deep Learning techniques have achieved remarkable results in many domains. Often, training deep learning models requires large datasets, which may require sensitive information to be uploaded to the cloud to accelerate training. To adequately protect sensitive information, we propose distributed layer-partitioned train…

GAPA method provides efficient uncertainty quantification for pretrained networks.

Training neural networks under a strict Lipschitz constraint is useful for provable adversarial robustness, generalization bounds, interpretable gradients, and Wasserstein distance estimation. By the composition property of Lipschitz functions, it suffices to ensure that each individual affine transformation or nonline…

Pruning neural networks adds differential privacy noise, preserving data utility.

We present a novel factor analysis method that can be applied to the discovery of common factors shared among trajectories in multivariate time series data. These factors satisfy a precedence-ordering property: certain factors are recruited only after some other factors are activated. Precedence-ordering arise in appli…

Active inference framework improves $U$-statistic estimation efficiency.

New method preserves privacy while improving machine learning accuracy.

Framework tackles class imbalance and noisy labels in active learning.

Sensitive inferences and user re-identification are major threats to privacy when raw sensor data from wearable or portable devices are shared with cloud-assisted applications. To mitigate these threats, we propose mechanisms to transform sensor data before sharing them with applications running on users' devices. Thes…

BiTAT improves neural network quantization for edge devices by focusing on weight dependencies and disentangling them.

Unified framework recovers exact input from SOM activation patterns.

In the last few years there has been a growing interest in Human Activity Recognition~(HAR) topic. Sensor-based HAR approaches, in particular, has been gaining more popularity owing to their privacy preserving nature. Furthermore, due to the widespread accessibility of the internet, a broad range of streaming-based app…

Study develops a new model for predicting individual mobility based on activity patterns.

An increasing number of sensors on mobile, Internet of things (IoT), and wearable devices generate time-series measurements of physical activities. Though access to the sensory data is critical to the success of many beneficial applications such as health monitoring or activity recognition, a wide range of potentially …

ADS filters data points for efficient batch active learning.

Standard neural network architectures are non-linear only by virtue of a simple element-wise activation function, making them both brittle and excessively large. In this paper, we consider methods for making the feed-forward layer more flexible while preserving its basic structure. We develop simple drop-in replacement…

FAQ efficiently evaluates LLMs with statistical guarantees using adaptive query selection.

Study efficient interactive learning for structured outputs with reliable computation.

Union of Subspaces (UoS) is a popular model to describe the underlying low-dimensional structure of data. The fine details of UoS structure can be described in terms of canonical angles (also known as principal angles) between subspaces, which is a well-known characterization for relative subspace positions. In this pa…

Data-aware activation function customization reduces neural network error.

This paper proposes grid cells encode position via a conformal isometric embedding of 2D physical space.

PPAT uses predictions to improve risk estimation in active testing.

Neural networks are known to be vulnerable to adversarial examples. Carefully chosen perturbations to real images, while imperceptible to humans, induce misclassification and threaten the reliability of deep learning systems in the wild. To guard against adversarial examples, we take inspiration from game theory and ca…

Before training a neural net, a classic rule of thumb is to randomly initialize the weights so the variance of activations is preserved across layers. This is traditionally interpreted using the total variance due to randomness in both weights \emph{and} samples. Alternatively, one can interpret the rule of thumb as pr…

We introduce a conceptually simple and scalable framework for continual learning domains where tasks are learned sequentially. Our method is constant in the number of parameters and is designed to preserve performance on previously encountered tasks while accelerating learning progress on subsequent problems. This is a…

A major challenge in computational chemistry is the generation of novel molecular structures with desirable pharmacological and physiochemical properties. In this work, we investigate the potential use of autoencoder, a deep learning methodology, for de novo molecular design. Various generative autoencoders were used t…

New method improves ResNet performance without batch normalization.

Post-hoc uncertainty quantification improves on pre-trained neural networks without underfitting.

Improved fraud detection in finance with quantum-enhanced federated learning.

Classification algorithms aim to predict an unknown label (e.g., a quality class) for a new instance (e.g., a product). Therefore, training samples (instances and labels) are used to deduct classification hypotheses. Often, it is relatively easy to capture instances but the acquisition of the corresponding labels remai…

It has been noted in existing literature that over-parameterization in ReLU networks generally improves performance. While there could be several factors involved behind this, we prove some desirable theoretical properties at initialization which may be enjoyed by ReLU networks. Specifically, it is known that He initia…

Paper connects MoE and self-attention, proposing active-attention.

We consider a few quantities that characterize trading on a stock market in a fixed time interval: logarithmic returns, volatility, trading activity (i.e., the number of transactions), and volume traded. We search for the power-law cross-correlations among these quantities aggregated over different time units from 1 mi…

A new deep learning framework selects representative samples for unsupervised learning.

We propose new symplectic networks (SympNets) for identifying Hamiltonian systems from data based on a composition of linear, activation and gradient modules. In particular, we define two classes of SympNets: the LA-SympNets composed of linear and activation modules, and the G-SympNets composed of gradient modules. Cor…

Theory explains deep nonlinear networks' plateaus and transitions.

This paper considers the scenario that multiple data owners wish to apply a machine learning method over the combined dataset of all owners to obtain the best possible learning output but do not want to share the local datasets owing to privacy concerns. We design systems for the scenario that the stochastic gradient d…