However, many use cases of neural network verification require solving the inverse problem, or over-approximating the set of inputs that lead to certain outputs.

The ability to encode informative functional beliefs in BNN priors can significantly reduce the bias and uncertainty of the posterior predictive, especially in regions of input space sparsely coveredbytraining data[27].

However, many use cases of neural network verification require solving the inverse problem, or over-approximating the set of inputs that lead to certain outputs.

-- Control barrier functions (CBFs) are a powerful tool for the constrained control of nonlinear systems; however, the majority of results in the literature focus on systems subject to a single CBF constraint, making it challenging to synthesize provably safe controllers that handle multiple state constraints. This paper presents a framework for constrained control of nonlinear systems subject to box constraints on the systems' vector-valued outputs using multiple CBFs. Our results illustrate that when the output has a vector relative degree, the CBF constraints encoding these box constraints are compatible, and the resulting optimization-based controller is locally Lipschitz continuous and admits a closed-form expression. Additional results are presented to characterize the degradation of nominal tracking objectives in the presence of safety constraints. Simulations of a planar quadrotor are presented to demonstrate the efficacy of the proposed framework.

Work done while at Harvard University.

The paper converts the constrained optimization problem to min-max optimization using Lagrangian function. To show the efficacy of the model, three experiments are conducted in 5.1 SRL, 5.2 NER and 5.3 Fine grained entity typing. The paper brings in a structured way of training with output constraints. However, I am not sure how much gain this model has on top of fixed weight on constraints (Metha et al 2018 & Diligenti et al 2017) with the provided experiments. Also while the experiments seem convincing as itself, it is hard to see how much significance this work brings in as the baselines significantly differ with related work. Also, it would give a better picture of this method if the paper could provide more analysis: an analysis on convergence, an analysis on experiment results on why more labeled data sometimes hurt, etc. [originality ] 1. The full Lagrangian expression and linking the output constraint to the model parameter and optimizing them with subgradient seems novel. However, how exactly the authors formulate f(w) is unclear to me. Is it just following the way Diligenti 2017 does it?

Prioritization of privacy can limit the effectiveness of these interactions, as achieving certain goals necessitates sharing private data. Traditionally, addressing this challenge has involved either seeking trusted intermediaries or constructing cryptographic protocols that restrict how much data is revealed, such as multi-party computations or zero-knowledge proofs. While significant advances have been made in scaling cryptographic approaches, they remain limited in terms of the size and complexity of applications they can be used for. In this paper, we argue that capable machine learning models can fulfill the role of a trusted third party, thus enabling secure computations for applications that were previously infeasible. In particular, we describe Trusted Capable Model Environments (TCMEs) as an alternative approach for scaling secure computation, where capable machine learning model(s) interact under input/output constraints, with explicit information flow control and explicit statelessness. This approach aims to achieve a balance between privacy and computational efficiency, enabling private inference where classical cryptographic solutions are currently infeasible. We describe a number of use cases that are enabled by TCME, and show that even some simple classic cryptographic problems can already be solved with TCME. Finally, we outline current limitations and discuss the path forward in implementing them. In this paper we contend that recent advancements in machine learning enable a new paradigm for private inference. Fundamentally, the need for many cryptographic primitives stems from the fact that we don't have trusted third parties, thus requiring mutually untrusted participants to interact in a way that avoids revealing their data to each other but where they can nevertheless agree on a result.

Each specification defines the expected model output for a given input space ( 2.1). The increasing adoption of neural networks in learning-augmented systems highlights the importance Specifically, researchers have relied on their domain knowledge of model safety and robustness, particularly and intuition about individual applications to manually in safety-critical domains. Despite progress in create specifications. For instance, in adaptive video streaming, the formal verification of neural networks, current where a neural network is employed to determine the practices require users to manually define model bitrate for the next video chunk based on recent network specifications--properties that dictate expected conditions, Eliyahu et al. (2021) define a specification as, model behavior in various scenarios. This manual "[if video] chunks were downloaded quickly (more quickly process, however, is prone to human error, limited than it takes to play a chunk), the DNN should eventually in scope, and time-consuming. In this paper, not choose the worst resolution." Similar manual specifications we introduce AutoSpec, the first framework to are devised for other learning-augmented systems, e.g., automatically generate comprehensive and accurate database indexes (Tan et al., 2021), memory allocators (Wei specifications for neural networks in learningaugmented et al., 2023), and job schedulers (Wu et al., 2022).

Most work on the formal verification of neural networks has focused on bounding the set of outputs that correspond to a given set of inputs (for example, bounded perturbations of a nominal input). However, many use cases of neural network verification require solving the inverse problem, or over-approximating the set of inputs that lead to certain outputs. We present the INVPROP algorithm for verifying properties over the preimage of a linearly constrained output set, which can be combined with branch-and-bound to increase precision. Contrary to other approaches, our efficient algorithm is GPU-accelerated and does not require a linear programming solver. We demonstrate our algorithm for identifying safe control regions for a dynamical system via backward reachability analysis, verifying adversarial robustness, and detecting out-of-distribution inputs to a neural network. Our results show that in certain settings, we find over-approximations over 2500 tighter than prior work while being 2.5 faster. By strengthening robustness verification with output constraints, we consistently verify more properties than the previous state-of-the-art on multiple benchmarks, including a large model with 167k neurons in VNN-COMP 2023. Our algorithm has been incorporated into the α,β-CROWN verifier, available at https://abcrown.org.

In advanced manufacturing, strict safety guarantees are required to allow humans and robots to work together in a shared workspace. One of the challenges in this application field is the variety and unpredictability of human behavior, leading to potential dangers for human coworkers. This paper presents a novel control framework by adopting safety-critical control and uncertainty estimation for human-robot collaboration. Additionally, to select the shortest path during collaboration, a novel quadratic penalty method is presented. The innovation of the proposed approach is that the proposed controller will prevent the robot from violating any safety constraints even in cases where humans move accidentally in a collaboration task. This is implemented by the combination of a time-varying integral barrier Lyapunov function (TVIBLF) and an adaptive exponential control barrier function (AECBF) to achieve a flexible mode switch between path tracking and collision avoidance with guaranteed closed-loop system stability. The performance of our approach is demonstrated in simulation studies on a 7-DOF robot manipulator. Additionally, a comparison between the tasks involving static and dynamic targets is provided.