Stochastic-YOLO: Efficient Probabilistic Object Detection under Dataset Shifts

Azevedo, Tiago; de Jong, René; Mattina, Matthew; Maji, Partha

doi:10.48550/arxiv.2009.02967

Cited by 7 publications

(16 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We consider different ways to apply DropBlock in YOLO models and show how it can improve the uncertainty modeling, generalization performance, and calibration capabilities of the YOLO models. We compare the proposed MC DropBlock with various baselines including the ones considering drop block only at training time (training time dropBlock) or only at inference or test time (inference time dropBlock), and with existing approaches such as Dropout based [37] and Gaussian Yolo models [38]. All the experiments are trained with NVIDIA V100 GPUs.…”

Section: Methodsmentioning

confidence: 99%

Monte Carlo DropBlock for Modelling Uncertainty in Object Detection

Deepshikha,

Yelleni,

Srijith

et al. 2021

Preprint

View full text Add to dashboard Cite

With the advancements made in deep learning, computer vision problems like object detection and segmentation have seen a great improvement in performance. However, in many real-world applications such as autonomous driving vehicles, the risk associated with incorrect predictions of objects is very high. Standard deep learning models for object detection such as YOLO models are often overconfident in their predictions and do not take into account the uncertainty in predictions on out-of-distribution data. In this work, we propose an efficient and effective approach to model uncertainty in object detection and segmentation tasks using Monte-Carlo DropBlock (MC-DropBlock) based inference. The proposed approach applies drop-block during training time and test time on the convolutional layer of the deep learning models such as YOLO. We show that this leads to a Bayesian convolutional neural network capable of capturing the epistemic uncertainty in the model. Additionally, we capture the aleatoric uncertainty using a Gaussian likelihood. We demonstrate the effectiveness of the proposed approach on modeling uncertainty in object detection and segmentation tasks using out-of-distribution experiments. Experimental results show that MC-DropBlock improves the generalization, calibration, and uncertainty modeling capabilities of YOLO models in object detection and segmentation.

show abstract

Section: Methodsmentioning

confidence: 99%

Monte Carlo DropBlock for Modelling Uncertainty in Object Detection

Deepshikha,

Yelleni,

Srijith

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…For SMCDO we use Wide-ResNet-20-k with a widening factor k = 3 and M = 3 samples. Dropout is added to approximately the second halve of the model (Convolution layer 13 and deeper), since previous experiments have shown that this configuration yields a good performance to latency trade-off [14]. Dropout layer are always positioned before a convolution layer.…”

Section: A Experimental Setup For Cifar-10mentioning

confidence: 99%

On Efficient Uncertainty Estimation for Resource-Constrained Mobile Applications

Rock¹,

Azevedo²,

Jong³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Deep neural networks have shown great success in prediction quality while reliable and robust uncertainty estimation remains a challenge. Predictive uncertainty supplements model predictions and enables improved functionality of downstream tasks including embedded and mobile applications, such as virtual reality, augmented reality, sensor fusion, and perception. These applications often require a compromise in complexity to obtain uncertainty estimates due to very limited memory and compute resources. We tackle this problem by building upon Monte Carlo Dropout (MCDO) models using the Axolotl framework; specifically, we diversify sampled subnetworks, leverage dropout patterns, and use a branching technique to improve predictive performance while maintaining fast computations. We conduct experiments on (1) a multi-class classification task using the CIFAR10 dataset, and (2) a more complex human body segmentation task. Our results show the effectiveness of our approach by reaching close to Deep Ensemble prediction quality and uncertainty estimation, while still achieving faster inference on resource-limited mobile platforms.

show abstract

“…Although the design can achieve a high throughput, the restriction of the nonlinear activation functions limits generality for different application scenarios. In [5], the authors propose software-based intermediate-layer caching (IC), evaluated in last layer BNNs.…”

Section: Related Work a Field Programmable Gate Array-based Acceleratorsmentioning

confidence: 99%

“…To further improve the overall hardware performance, we propose a hardware implementation of IC technique [5] to decrease the required compute and the number of memory accesses. An example of using IC is illustrated in Figure 4, where the NN contains two layers and it only requires the user to apply the dropout mask and run the last layer S times when the partial Bayesian technique is applied.…”

Section: Intermediate-layer Caching (Ic)mentioning

confidence: 99%

“…In recent years, neural networks (NNs) have demonstrated their outstanding performance in a variety of applications ranging from image classification [1] or segmentation [2] to human action recognition [3]. However, one of the main drawbacks in standard NNs is that they are not able to capture the model uncertainty which is crucial for many safetycritical applications such as healthcare [4] or autonomous vehicles [5]. In contrast to standard NNs, Bayesian neural networks (BNNs) [6], which adopt Bayesian inference to provide a principled uncertainty estimation, have become more popular in these applications.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

High-Performance FPGA-based Accelerator for Bayesian Neural Networks

Fan

Ferianc

Rodrigues

et al. 2021

Preprint

View full text Add to dashboard Cite

Neural networks (NNs) have demonstrated their potential in a wide range of applications such as image recognition, decision making or recommendation systems. However, standard NNs are unable to capture their model uncertainty which is crucial for many safety-critical applications including healthcare and autonomous vehicles. In comparison, Bayesian neural networks (BNNs) are able to express uncertainty in their prediction via a mathematical grounding. Nevertheless, BNNs have not been as widely used in industrial practice, mainly because of their expensive computational cost and limited hardware performance. This work proposes a novel FPGAbased hardware architecture to accelerate BNNs inferred through Monte Carlo Dropout. Compared with other state-of-the-art BNN accelerators, the proposed accelerator can achieve up to 4 times higher energy efficiency and 9 times better compute efficiency. Considering partial Bayesian inference, an automatic framework is proposed, which explores the trade-off between hardware and algorithmic performance. Extensive experiments are conducted to demonstrate that our proposed framework can effectively find the optimal points in the design space.• A novel hardware architecture with an intermediate-layer caching technique to accelerate Bayesian neural networks

show abstract

Stochastic-YOLO: Efficient Probabilistic Object Detection under Dataset Shifts

Cited by 7 publications

References 14 publications

Monte Carlo DropBlock for Modelling Uncertainty in Object Detection

Monte Carlo DropBlock for Modelling Uncertainty in Object Detection

On Efficient Uncertainty Estimation for Resource-Constrained Mobile Applications

High-Performance FPGA-based Accelerator for Bayesian Neural Networks

Contact Info

Product

Resources

About