arXiv subjects

Richard Klein

Publications and source records attributed to Richard Klein.

22 records · Page 2Linked to original sources

arXiv2021

Confident in the Crowd: Bayesian Inference to Improve Data Labelling in Crowdsourcing

With the increased interest in machine learning and big data problems, the need for large amounts of labelled data has also grown. However, it is often infeasible to get experts to label all of this data, which leads many practitioners to crowdsourcing solutions. In this paper, we present new techniques to improve the quality of the labels while attempting to reduce the cost. The naive approach to assigning labels is to adopt a majority vote method, however, in the context of data labelling, this is not always ideal as data labellers are not equally reliable. One might, instead, give higher priority to certain labellers through some kind of weighted vote based on past performance. This paper investigates the use of more sophisticated methods, such as Bayesian inference, to measure the performance of the labellers as well as the confidence of each label. The methods we propose follow an iterative improvement algorithm which attempts to use the least amount of workers necessary to achieve the desired confidence in the inferred label. This paper explores simulated binary classification problems with simulated workers and questions to test the proposed methods. Our methods outperform the standard voting methods in both cost and accuracy while maintaining higher reliability when there is disagreement within the crowd.

cs.LG↗

arXiv2021

Explicit homography estimation improves contrastive self-supervised learning

The typical contrastive self-supervised algorithm uses a similarity measure in latent space as the supervision signal by contrasting positive and negative images directly or indirectly. Although the utility of self-supervised algorithms has improved recently, there are still bottlenecks hindering their widespread use, such as the compute needed. In this paper, we propose a module that serves as an additional objective in the self-supervised contrastive learning paradigm. We show how the inclusion of this module to regress the parameters of an affine transformation or homography, in addition to the original contrastive objective, improves both performance and learning speed. Importantly, we ensure that this module does not enforce invariance to the various components of the affine transform, as this is not always ideal. We demonstrate the effectiveness of the additional objective on two recent, popular self-supervised algorithms. We perform an extensive experimental analysis of the proposed method and show an improvement in performance for all considered datasets. Further, we find that although both the general homography and affine transformation are sufficient to improve performance and convergence, the affine transformation performs better in all cases.

cs.CV↗

arXiv2020

Quantisation and Pruning for Neural Network Compression and Regularisation

Deep neural networks are typically too computationally expensive to run in real-time on consumer-grade hardware and low-powered devices. In this paper, we investigate reducing the computational and memory requirements of neural networks through network pruning and quantisation. We examine their efficacy on large networks like AlexNet compared to recent compact architectures: ShuffleNet and MobileNet. Our results show that pruning and quantisation compresses these networks to less than half their original size and improves their efficiency, particularly on MobileNet with a 7x speedup. We also demonstrate that pruning, in addition to reducing the number of parameters in a network, can aid in the correction of overfitting.

cs.LG↗

arXiv2014

Star Cluster Formation in Turbulent, Magnetized Dense Clumps with Radiative and Outflow Feedback

We present three Orion simulations of star cluster formation in a 1000 Msun, turbulent molecular cloud clump, including the effects of radiative transfer, protostellar outflows, and magnetic fields. Our simulations all use self-consistent turbulent initial conditions and vary the mean mass-to-flux ratio relative to the critical value over 2, 10, and infinity to gauge the influence of magnetic fields on star cluster formation. We find, in good agreement with previous studies, that magnetic fields of typically observed strengths lower the star formation rate by a factor of 2.4 and reduce the amount of fragmentation by a factor of 2 relative to the zero-field case. We also find that the field increases the characteristic sink particle mass, again by a factor of 2.4. The magnetic field also increases the degree of clustering in our simulations, such that the maximum stellar densities in the strong field case are higher than the others by again a factor of 2. This clustering tends to encourage the formation of multiple systems, which are more common in the rad-MHD runs than the rad-hydro run. The companion frequency in our simulations is consistent with observations of multiplicity in Class I sources, particularly for the strong field case. Finally, we find evidence of primordial mass segregation in our simulations reminiscent of that observed in star clusters like the Orion Nebula Cluster.

astro-ph.GA↗