ShardNet: Training Neural Controllers with Hard, Non-Convex Constraints

• URL: http://arxiv.org/abs/2606.30935v1
• Date: Mon, 29 Jun 2026 21:38:44 GMT
• Title: ShardNet: Training Neural Controllers with Hard, Non-Convex Constraints
• Abstract summary: We introduce ShardNet, a neural network architecture that enforces unions of constraints by construction.<n>We show that ShardNet policies maintain 100% safety on verified sets and achieves significantly lower objective loss compared to existing formal methods.
• Score: 2.1198879079315573
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: While neural network control policies are powerful, their deployment on safety critical systems depends on ensuring that they obey strict constraints. Existing work often treats safety as a metric to optimize for, which competes with other performance objectives, if training converges at all. Instead, we introduce ShardNet, a neural network architecture that strictly enforces unions of polyhedral constraints by construction, using a differentiable projection layer parameterized by a classification network. The key insight is to embed safety into the neural network's structure, allowing performance to be optimized independently because formal safety guarantees are always given. In contrast with existing neural architectures that can only enforce simple convex constraints, ShardNet enables the first safe-by-construction synthesis of forward-invariant neural network controllers on closed-loop systems where safety constraints are expressed as nonconvex unions of polyhedras or learned value function level sets. To support this, we also introduce a technique to verify and train such value functions correctly as rectified linear unit (ReLU) networks, which has not previously been possible. On double integrator benchmarks drawn from the literature, ShardNet policies maintain 100% safety on verified sets and achieves significantly lower objective loss compared to existing formal methods. Furthermore, our value function training technique also produces safe sets more than 3 times larger than existing verification approaches.

Related papers

• PoSafeNet: Safe Learning with Poset-Structured Neural Nets [49.854863600271614]
  existing approaches often enforce multiple safety constraints uniformly or via fixed priority orders, leading to infeasibility and brittle behavior.<n>We formalize this setting as poset-structured safety, modeling safety constraints as a partially ordered set and treating safety composition as a structural property of the policy class.<n>Building on this formulation, we propose PoSafeNet, a differentiable neural safety layer that enforces safety via sequential closed-form projection.
  arXiv  Detail & Related papers  (2026-01-29T22:03:32Z)
• Provably-Safe Neural Network Training Using Hybrid Zonotope Reachability Analysis [2.1198879079315573]
  It is difficult to enforce constraints on neural networks in safety-critical control applications.<n>We propose a method that can encourage exact image of a non- hybrid input set for a neural network with rectified linear unit (ReLU) nonlinearities.
  arXiv  Detail & Related papers  (2025-01-22T17:13:48Z)
• HardNet: Hard-Constrained Neural Networks with Universal Approximation Guarantees [4.814150850377003]
  HardNet is a framework for constructing neural networks that inherently satisfy hard constraints.<n>We show that HardNet retains neural networks' universal approximation capabilities.
  arXiv  Detail & Related papers  (2024-10-14T17:59:24Z)
• Probabilistic Reach-Avoid for Bayesian Neural Networks [71.67052234622781]
  We show that an optimal synthesis algorithm can provide more than a four-fold increase in the number of certifiable states. The algorithm is able to provide more than a three-fold increase in the average guaranteed reach-avoid probability.
  arXiv  Detail & Related papers  (2023-10-03T10:52:21Z)
• Robust Training and Verification of Implicit Neural Networks: A Non-Euclidean Contractive Approach [64.23331120621118]
  This paper proposes a theoretical and computational framework for training and robustness verification of implicit neural networks. We introduce a related embedded network and show that the embedded network can be used to provide an $ell_infty$-norm box over-approximation of the reachable sets of the original network. We apply our algorithms to train implicit neural networks on the MNIST dataset and compare the robustness of our models with the models trained via existing approaches in the literature.
  arXiv  Detail & Related papers  (2022-08-08T03:13:24Z)
• BarrierNet: A Safety-Guaranteed Layer for Neural Networks [50.86816322277293]
  BarrierNet allows the safety constraints of a neural controller be adaptable to changing environments. We evaluate them on a series of control problems such as traffic merging and robot navigations in 2D and 3D space.
  arXiv  Detail & Related papers  (2021-11-22T15:38:11Z)
• OVERT: An Algorithm for Safety Verification of Neural Network Control Policies for Nonlinear Systems [31.3812947670948]
  We present OVERT: a sound algorithm for safety verification of neural network control policies. The central concept of OVERT is to abstract nonlinear functions with a set of optimally tight piecewise linear bounds. Overt compares favorably to existing methods both in time and in tightness of the reachable set.
  arXiv  Detail & Related papers  (2021-08-03T00:41:27Z)
• Self-Repairing Neural Networks: Provable Safety for Deep Networks via Dynamic Repair [16.208330991060976]
  We propose a way to construct neural network classifiers that dynamically repair violations of non-relational safety constraints. Our approach is based on a novel self-repairing layer, which provably yields safe outputs. We show that our approach can be implemented using vectorized computations that execute efficiently on a GPU.
  arXiv  Detail & Related papers  (2021-07-23T20:08:52Z)
• Constrained Feedforward Neural Network Training via Reachability Analysis [0.0]
  It remains an open challenge to train a neural network to obey safety constraints. This work proposes a constrained method to simultaneously train and verify a feedforward neural network with rectified linear unit (ReLU) nonlinearities.
  arXiv  Detail & Related papers  (2021-07-16T04:03:01Z)
• Safe RAN control: A Symbolic Reinforcement Learning Approach [62.997667081978825]
  We present a Symbolic Reinforcement Learning (SRL) based architecture for safety control of Radio Access Network (RAN) applications. We provide a purely automated procedure in which a user can specify high-level logical safety specifications for a given cellular network topology. We introduce a user interface (UI) developed to help a user set intent specifications to the system, and inspect the difference in agent proposed actions.
  arXiv  Detail & Related papers  (2021-06-03T16:45:40Z)
• Enforcing robust control guarantees within neural network policies [76.00287474159973]
  We propose a generic nonlinear control policy class, parameterized by neural networks, that enforces the same provable robustness criteria as robust control. We demonstrate the power of this approach on several domains, improving in average-case performance over existing robust control methods and in worst-case stability over (non-robust) deep RL methods.
  arXiv  Detail & Related papers  (2020-11-16T17:14:59Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.