Paper presents an ADMM-based approach to efficiently integrate quadratic programming layers into neural networks.
problem Integrating quadratic programs into neural networks for optimization.
method An ADMM-based network layer architecture for solving quadratic programs efficiently.
result The ADMM layer is approximately an order of magnitude faster than existing methods for medium scaled problems.
Integrates fairness guarantees into deep learning models.
problem Ensuring fairness in deep learning models.
method Integrates a differentiable fairness layer into neural models and uses an online primal-dual algorithm for provable fairness guarantees.
result Guarantees a chosen notion of output parity in deep learning models.
We inductively define layers of colorings of knot and knotted surface diagrams using ternary quasigroups. Homological invariants from such systems of colorings use shorter differentials and of higher degree than the standard homology differentials, and give access to typically more complex homology groups.
Normalization layers improve the accuracy of Differentially Private training of deep neural networks.
problem Reduced accuracy in deep neural networks with Differentially Private training.
method Proposed a novel method for integrating batch normalization with Differentially Private Stochastic Gradient Descent (DPSGD) without additional privacy loss.
result Training deeper networks with better utility-privacy trade-off is possible.
Recent work has shown how to embed differentiable optimization problems (that is, problems whose solutions can be backpropagated through) as layers within deep learning architectures. This method provides a useful inductive bias for certain problems, but existing software for differentiable optimization layers is rigid…
Pruning neural networks adds differential privacy noise, preserving data utility.
problem Achieving differential privacy in neural networks without sacrificing data utility.
method Proving equivalence between pruning and adding differential privacy noise to hidden-layer activations.
result Pruning can be a more effective alternative to adding differential privacy noise for neural networks.
BPQP improves efficiency of differentiable optimization layers for deep learning.
problem Efficiency in differentiating optimization problems for deep learning models.
method Reformulates optimization problems as quadratic programming, enabling efficient gradient calculation.
result Significant improvement in efficiency (order of magnitude faster) compared to other differentiable optimization layers.
This paper presents OptNet, a network architecture that integrates optimization problems (here, specifically in the form of quadratic programs) as individual layers in larger end-to-end trainable deep networks. These layers encode constraints and complex dependencies between the hidden states that traditional convoluti…
For one-hidden-layer ReLU networks, we prove that all differentiable local minima are global inside differentiable regions. We give the locations and losses of differentiable local minima, and show that these local minima can be isolated points or continuous hyperplanes, depending on an interplay between data, activati…
The paper extends mean field results to three-layer neural networks using SGD.
problem Understanding the dynamics of training three-layer neural networks with SGD.
method Extending mean field results from two-layer networks to three-layer networks with two hidden layers, using non-linear partial differential equations.
result The distributions of weights in the two hidden layers are independent.
Differentiable Masking reveals how neural models make decisions across layers.
problem Intractable and expensive approximate search for input relevance in deep models.
method Differentiable Masking learns to mask inputs while maintaining differentiability.
result Reveals how decisions are formed across network layers in BERT models.
Multi-layered representation is believed to be the key ingredient of deep neural networks especially in cognitive tasks like computer vision. While non-differentiable models such as gradient boosting decision trees (GBDTs) are the dominant methods for modeling discrete or tabular data, they are hard to incorporate with…
Differentiable cutting-plane layers solve parametric mixed-integer linear optimization problems.
problem Solving parametric mixed-integer linear optimization problems with changing data.
method Introducing cutting-plane layers (CPLs) for differentiable cutting-plane generation.
result The algorithm computes solutions with low integrality gaps and generalizes to unseen instances.
We consider deep linear networks with arbitrary convex differentiable loss. We provide a short and elementary proof of the fact that all local minima are global minima if the hidden layers are either 1) at least as wide as the input layer, or 2) at least as wide as the output layer. This result is the strongest possibl…
We introduce a technique based on the singular vector canonical correlation analysis (SVCCA) for measuring the generality of neural network layers across a continuously-parametrized set of tasks. We illustrate this method by studying generality in neural networks trained to solve parametrized boundary value problems ba…
DSA efficiently allocates sparsity across layers for budgeted pruning.
problem Efficiently distributing resources (sparsity) across layers in pruning under resource constraints.
method DSA uses differentiable pruning to find continuous layer-wise pruning ratios via gradient-based optimization.
result DSA achieves superior performance and significantly reduces the time cost of pruning.
Improved differentially private deep learning with group-wise clipping techniques.
problem Efficiency and privacy trade-offs in deep learning models.
method Group-wise clipping techniques (per-layer and per-device) to reduce compute time and memory overhead.
result Private learning with group-wise clipping achieves similar or better performance than non-private learning with less wall time.
Embedding layers are commonly used to map discrete symbols into continuous embedding vectors that reflect their semantic meanings. Despite their effectiveness, the number of parameters in an embedding layer increases linearly with the number of symbols and poses a critical challenge on memory and storage constraints. I…
Paper introduces probabilistic approach to CO layers in ML.
problem Challenges in training CO layers with stochastic gradient descent.
method Probabilistic perspective, approximate differentiation, structured losses.
result InferOpt.jl package for training CO layers.
NDM incorporates geometric structure into neural networks for better optimization and interpretability.
problem Efficient and interpretable deep learning architectures.
method NDM is a neural network architecture that explicitly incorporates geometric structure into its design, using a Coordinate Layer, Geometric Layer, and Evolution Layer.
result NDM provides intrinsic regularization, enhancing generalization and robustness.
Neural networks and tree ensembles are state-of-the-art learners, each with its unique statistical and computational advantages. We aim to combine these advantages by introducing a new layer for neural networks, composed of an ensemble of differentiable decision trees (a.k.a. soft trees). While differentiable trees dem…
Physarum-inspired solver speeds up LP solving in neural networks.
problem Efficiently solving linear programming problems within neural networks.
method Inspired by Physarum dynamics, a differentiable solver for LPs.
result Solver outperforms existing methods in video segmentation and few-shot learning.
MSLs use parallelizable root-finding for efficient ODE and PDE solutions.
problem Efficiently solving initial value problems for ODEs and PDEs.
method Leveraging time-parallel methods, MSLs use parallelizable root-finding algorithms.
result MSLs offer significant speedups in NFEs and inference time.
We establish the first benchmark for federated learning with differential privacy in ASR, achieving competitive performance.
problem Challenges in training large transformer models for ASR in federated learning with differential privacy.
method Per-layer clipping and layer-wise gradient normalization to mitigate gradient heterogeneity.
result FL with DP is viable in ASR with strong privacy guarantees, achieving competitive performance.
We prove that a single-layer neural network trained with the Q-learning algorithm converges in distribution to a random ordinary differential equation as the size of the model and the number of training steps become large. Analysis of the limit differential equation shows that it has a unique stationary solution which …
We propose a novel deep learning paradigm of differential flows that learn a stochastic differential equation transformations of inputs prior to a standard classification or regression function. The key property of differential Gaussian processes is the warping of inputs through infinitely deep, but infinitesimal, diff…
In deep neural networks, the spectral norm of the Jacobian of a layer bounds the factor by which the norm of a signal changes during forward/backward propagation. Spectral norm regularizations have been shown to improve generalization, robustness and optimization of deep learning methods. Existing methods to compute th…
VPP learns joint policies for multi-agent RL through interactions.
problem Learning effective joint policies for multi-agent reinforcement learning.
method VPP integrates variational inference into policy layers for efficient sampling and differentiability.
result VPP outperforms previous methods on large-scale multi-agent tasks.
Current deep learning models are mostly build upon neural networks, i.e., multiple layers of parameterized differentiable nonlinear modules that can be trained by backpropagation. In this paper, we explore the possibility of building deep models based on non-differentiable modules. We conjecture that the mystery behind…
A new layer learns abstract relations from graph structure using finite-state automata.
problem Learning abstract relations from graph structure for program analysis.
method Relaxing the problem into learning finite-state automata policies on a graph-based POMDP and training these policies using implicit differentiation.
result GFSA layer finds shortcuts in grid-world graphs and reproduces simple static analyses on Python programs.
Recently, several studies proposed methods to utilize some classes of optimization problems in designing deep neural networks to encode constraints that conventional layers cannot capture. However, these methods are still in their infancy and require special treatments, such as analyzing the KKT condition, for deriving…
Lie groupoid equivariant neural networks are a new type of neural network.
problem Designing neural networks that respect the structure of Lie groupoids.
method Introducing Lie groupoid equivariant convolutions and layers, and showing their equivalence to Lie algebroid-equivariant networks.
result Lie groupoid equivariant neural networks are equivalent to certain Lie algebroid-equivariant networks.
The paper addresses differentially private learning for neural networks, focusing on risk bounds and algorithm feasibility.
problem Achieving differentially private learning for neural networks with theoretical guarantees.
method Developed algorithms and theoretical analysis for differentially private stochastic optimization of neural networks.
result Established theoretical bounds for excess population risk in differentially private learning of neural networks.
SnareNet adds repair layers to neural networks to ensure outputs meet physical constraints.
problem Unconstrained neural network predictions violate physical or safety requirements.
method SnareNet appends a differentiable repair layer that navigates constraints to produce feasible outputs.
result SnareNet consistently improves objective quality while satisfying constraints more reliably.
DNArch learns CNN architectures by backpropagation.
problem Discovering optimal CNN architectures.
method Differentiable Neural Architectures (DNArch) learns CNN architectures by backpropagation, controlling kernel sizes, channels, downsampling positions, and depth.
result DNArch finds performant CNN architectures across various tasks.
The Brylinski beta function is extended for coaxial layers on submanifolds.
problem Extending the Brylinski beta function to coaxial layers on submanifolds.
method Analytic continuation and computation of residues for the function.
result The Brylinski beta function has an analytic continuation with simple poles.
TF Boosted Trees (TFBT) is a new open-sourced frame-work for the distributed training of gradient boosted trees. It is based on TensorFlow, and its distinguishing features include a novel architecture, automatic loss differentiation, layer-by-layer boosting that results in smaller ensembles and faster prediction, princ…
Soft-Radial Projection solves gradient saturation in constrained deep learning.
problem Gradient saturation in deep learning models when integrating hard constraints.
method Introduces Soft-Radial Projection, a differentiable layer that maps predictions onto constraint boundaries without rank-deficient Jacobians.
result Improves convergence and solution quality over state-of-the-art methods.
We propose PLLay, a novel topological layer for general deep learning models based on persistence landscapes, in which we can efficiently exploit the underlying topological features of the input data structure. In this work, we show differentiability with respect to layer inputs, for a general persistent homology with …
The paper analyzes reinsurance strategies in a competitive multi-agent system.
problem Strategic interactions and competitive behavior in multi-layer reinsurance chains.
method Stochastic differential games and non-zero-sum game models to characterize strategic interactions. Dynamic programming and game theory to derive equilibrium strategies.
result Intensified competition reduces safety loadings in reinsurance contracts.
A new method for target propagation using iterative approximations converges fast and is more biologically plausible.
problem Improving target propagation methods for neural networks.
method Iterative approximate inverses and local auto-encoders.
result The method converges exponentially fast under certain conditions.
This paper extends geometric study of neural networks to non-differentiable layers and random walks.
problem Understanding the geometric properties of neural networks, especially those with non-differentiable activation functions.
method Singular Riemannian geometry approach to convolutional, residual, and recursive neural networks.
result Illustrated geometric findings with numerical experiments on image classification and thermodynamic problems.
Syndrome differentiation in Traditional Chinese Medicine (TCM) is the process of understanding and reasoning body condition, which is the essential step and premise of effective treatments. However, due to its complexity and lack of standardization, it is challenging to achieve. In this study, we consider each patient'…
PIVOT bridges Black-Scholes price and implied volatility spaces via a differentiable layer.
problem Lack of a differentiable interface between price and implied volatility spaces.
method Develops PIVOT, a differentiable layer that preserves LBR's forward pass and avoids backpropagation through branch logic, addressing singularity issues.
result PIVOT achieves high performance and accuracy, reducing price and implied volatility errors by up to 43.4% and 21.3% respectively.
A new neural network layer integrates graph learning into classification tasks.
problem Lack of relational information in standard deep learning architectures for label predictions.
method Derives backpropagation equations for a differentiable graph learning layer.
result Smooth label transitions, improved generalization, and robustness to adversarial attacks.
Algorithm learns which weights to share in deep multi-task learning.
problem Difficulty in deciding which weights to share between tasks in deep learning models.
method Combines natural evolution strategy and stochastic gradient descent to learn optimal weight sharing.
result Task-specific networks achieve lower test errors than existing methods on multi-task learning datasets.
Stochastic gradient descent (SGD) has achieved great success in training deep neural network, where the gradient is computed through back-propagation. However, the back-propagated values of different layers vary dramatically. This inconsistence of gradient magnitude across different layers renders optimization of deep …
A new framework for flexible neural network receptive fields.
problem Adaptive and flexible receptive fields in neural networks.
method Density-embedding layers that replace affine transformations with scalar products of input and density functions.
result Density-embedding layers can adaptively tune receptive fields and are computationally efficient.