arXiv subjects

Ben Lee

Publications and source records attributed to Ben Lee.

5 recordsLinked to original sources

arXiv2024

Dynamic Simultaneous Multithreaded Architecture

This paper presents the Dynamic Simultaneous Multi-threaded Architecture (DSMT). DSMT efficiently exe-cutes multiple threads from a single program on a SMT processor core. To accomplish this, threads are generated dynamically from a predictable flow of control and then executed speculatively. Data obtained during the single context non-speculative execution phase of DSMT is used as a hint to speculate the posterior behavior of multiple threads. DSMT employs simple mechanisms based on state bits that keep track of inter-thread dependencies in registers and memory, synchronize thread execution, and control recovery from misspeculation. Moreover, DSMT utilizes a novel greedy policy for choosing those sections of code which provide the highest performance based on their past execution history. The DSMT architecture was simulated with a new cycle-accurate, execution-driven simulator. Our simulation results show that DSMT has very good potential to improve SMT performance, even when only a single program is available. However, we found that dynamic thread behavior together with fre-quent misspeculation may also produce diminishing re-turns in performance. Therefore, the challenge is to max-imize the amount of thread-level parallelism that DSMT is capable of exploiting and at the same time reduce the fre-quency of misspeculations.

cs.AR↗

arXiv2024

Assessing survival models by interval testing with Poisson-binomial distributions

Selecting appropriate parametric survival models is often a pivotal part of a regulatory submission for new pharmaceutical products. With recent developments in complex survival approaches, the number of suitable models is increasing, making model selection more challenging. Common approaches to model selection include AIC, BIC, and expert opinion on survival extrapolation. However, these approaches primarily assess relative goodness-of-fit, providing limited insight into where, and to what extent, a fitted model is incompatible with the observed data. We propose evaluating survival models using Poisson-binomial distributions across specified time intervals. Two interval selection approaches, censor-defined intervals and 10 evenly-spaced intervals, are presented with worked examples. A simulation exercise, targeting two proposed test statistics across 12 standard scenarios (with different data maturity and patient numbers), demonstrated that for every scenario the empirical Type I error did not exceed the nominal 5% level. Our proposed model selection technique goes beyond classical approaches by highlighting time intervals where models perform poorly.

stat.ME↗

arXiv2024

Robust Inverse Graphics via Probabilistic Inference

How do we infer a 3D scene from a single image in the presence of corruptions like rain, snow or fog? Straightforward domain randomization relies on knowing the family of corruptions ahead of time. Here, we propose a Bayesian approach-dubbed robust inverse graphics (RIG)-that relies on a strong scene prior and an uninformative uniform corruption prior, making it applicable to a wide range of corruptions. Given a single image, RIG performs posterior inference jointly over the scene and the corruption. We demonstrate this idea by training a neural radiance field (NeRF) scene prior and using a secondary NeRF to represent the corruptions over which we place an uninformative prior. RIG, trained only on clean data, outperforms depth estimators and alternative NeRF approaches that perform point estimation instead of full inference. The results hold for a number of scene prior architectures based on normalizing flows and diffusion models. For the latter, we develop reconstruction-guidance with auxiliary latents (ReGAL)-a diffusion conditioning algorithm that is applicable in the presence of auxiliary latent variables such as the corruption. RIG demonstrates how scene priors can be used beyond generation tasks.

cs.CV↗

arXiv2022

ProbNeRF: Uncertainty-Aware Inference of 3D Shapes from 2D Images

The problem of inferring object shape from a single 2D image is underconstrained. Prior knowledge about what objects are plausible can help, but even given such prior knowledge there may still be uncertainty about the shapes of occluded parts of objects. Recently, conditional neural radiance field (NeRF) models have been developed that can learn to infer good point estimates of 3D models from single 2D images. The problem of inferring uncertainty estimates for these models has received less attention. In this work, we propose probabilistic NeRF (ProbNeRF), a model and inference strategy for learning probabilistic generative models of 3D objects' shapes and appearances, and for doing posterior inference to recover those properties from 2D images. ProbNeRF is trained as a variational autoencoder, but at test time we use Hamiltonian Monte Carlo (HMC) for inference. Given one or a few 2D images of an object (which may be partially occluded), ProbNeRF is able not only to accurately model the parts it sees, but also to propose realistic and diverse hypotheses about the parts it does not see. We show that key to the success of ProbNeRF are (i) a deterministic rendering scheme, (ii) an annealed-HMC strategy, (iii) a hypernetwork-based decoder architecture, and (iv) doing inference over a full set of NeRF weights, rather than just a low-dimensional code.

cs.CV↗

arXiv2007

Local acyclic fibrations and the de Rham complex

We reinterpret algebraic de Rham cohomology for a possibly singular complex variety X as sheaf cohomology in the site of smooth schemes over X with Voevodsky's h-topology. Our results extend to the algebraic de Rham complex as well. Our main technique is to extend Cech cohomology of hypercovers to arbitrary local acyclic fibrations of simplicial presheaves.

math.AG↗