Search

Generative AI holds the promise of enabling a range of sought-after capabilities and revolutionizing workflows in various consumer and enterprise verticals. However, putting a model in production involves much more than just generating an output. It involves ensuring the model is reliable, safe, performant and also adheres to the policy of operation in a particular domain. Guardrails as a necessity for models has evolved around the need to enforce appropriate behavior of models, especially when they are in production. In this paper, we use education as a use case, given its stringent requirements of the appropriateness of content in the domain, to demonstrate how a guardrail model can be trained and deployed in production. Specifically, we describe our experience in building a production-grade guardrail model for a K-12 educational platform. We begin by formulating the requirements for deployment to this sensitive domain. We then describe the training and benchmarking of our domain-specific guardrail model, which outperforms competing open- and closed- instruction-tuned models of similar and larger size, on proprietary education-related benchmarks and public benchmarks related to general aspects of safety. Finally, we detail the choices we made on architecture and the optimizations for deploying this service in production; these range across the stack from the hardware infrastructure to the serving layer to language model inference optimizations. We hope this paper will be instructive to other practitioners looking to create production-grade domain-specific services based on generative AI and large language models.

Gliomas are the most common malignant primary brain tumors in adults and one of the deadliest types of cancer. There are many challenges in treatment and monitoring due to the genetic diversity and high intrinsic heterogeneity in appearance, shape, histology, and treatment response. Treatments include surgery, radiation, and systemic therapies, with magnetic resonance imaging (MRI) playing a key role in treatment planning and post-treatment longitudinal assessment. The 2024 Brain Tumor Segmentation (BraTS) challenge on post-treatment glioma MRI will provide a community standard and benchmark for state-of-the-art automated segmentation models based on the largest expert-annotated post-treatment glioma MRI dataset. Challenge competitors will develop automated segmentation models to predict four distinct tumor sub-regions consisting of enhancing tissue (ET), surrounding non-enhancing T2/fluid-attenuated inversion recovery (FLAIR) hyperintensity (SNFH), non-enhancing tumor core (NETC), and resection cavity (RC). Models will be evaluated on separate validation and test datasets using standardized performance metrics utilized across the BraTS 2024 cluster of challenges, including lesion-wise Dice Similarity Coefficient and Hausdorff Distance. Models developed during this challenge will advance the field of automated MRI segmentation and contribute to their integration into clinical practice, ultimately enhancing patient care., Comment: 10 pages, 4 figures, 1 table

We perform a thorough analysis of de Sitter vacua in O(d,d) invariant cosmologies. Starting with a homogeneous and isotropic framework we examine conditions for the existence of such vacua, non-perturbative in \alpha' in both the string frame and the Einstein frame. We elucidate the nature of the instability in the string frame vacuum. For the Einstein frame, we demonstrate that the de Sitter vacuum cannot be eternal. We then extend our analysis to include Bianchi I universes where the O(d,d) symmetry includes scale factor exchange as well as scale factor duality. We show how the theory can be extended to the anisotropic case so that it admits de Sitter vacua, noting the crucial role played by the O(d,d) symmetry in satisfying any additional constraints., Comment: Paper updated with references added, most notably that of Rost and Nunez which includes details of additional dS solutions

The navigation of complex labyrinths with tens of rooms under visual partially observable state is typically addressed using recurrent deep reinforcement learning architectures. In this work, we show that navigation can be achieved through the emergent evolution of a simple Braitentberg-style heuristic that structures the interaction between agent and labyrinth, i.e. complex behaviour from simple heuristics. To do so, the approach of tangled program graphs is assumed in which programs cooperatively coevolve to develop a modular indexing scheme that only employs 0.8\% of the state space. We attribute this simplicity to several biases implicit in the representation, such as the use of pixel indexing as opposed to deploying a convolutional kernel or image processing operators.

Inspired by the method of smoothed asymptotics developed by Terence Tao, we introduce a new ultra-violet regularisation scheme for loop integrals in quantum field theory which we call $\eta$ regularisation. This allows us to reveal a surprising connection between the elimination of divergences in divergent series of powers and the preservation of gauge invariance in the regularisation of loop integrals in quantum field theory. In particular, we note that a method for regularising the series of natural numbers so that it converges to minus one twelfth inspires a regularisation scheme for non-abelian gauge theories coupled to Dirac fermions that preserves the Ward identity for the vacuum polarisation tensor. We also comment on a possible connection to Schwinger proper time integrals., Comment: 26 pages

Fabrication of superconducting nanomechanical resonators for quantum research, detectors and devices traditionally relies on a lithographic process, resulting in oscillators with sharp edges and a suspended length limited to a few 100 micrometres. We report a low-investment top-down approach to fabricating NbTi nanowire resonators with suspended lengths up to several millimetres and diameters down to 100 nanometres. The nanowires possess high critical currents and fields, making them a natural choice for magnetomotive actuation and sensing. This fabrication technique is independent of the substrate material, dimensions and layout and can readily be adapted to fabricate nanowire resonators from any metal or alloy with suitable ductility and yield strength. Our work thus opens access to a new class of nanomechanical devices with applications including microscopic and mesoscopic investigations of quantum fluids, detecting dark matter and fundamental materials research in one-dimensional superconductors in vacuum., Comment: 19 pages, 4 figures