Model learning from class imbalanced training data is a long-standing and significant challenge for machine learning. In particular, existing deep learning methods consider mostly either class balanced data or moderately imbalanced data in model training, and ignore the challenge of learning from significantly imbalanced training data. To address this problem, we formulate a class imbalanced deep learning model based on batch-wise incremental minority (sparsely sampled) class rectification by hard sample mining in majority (frequently sampled) classes during model training. This model is designed to minimise the dominant effect of majority classes by discovering sparsely sampled boundaries of minority classes in an iterative batch-wise learning process. To that end, we introduce a Class Rectification Loss (CRL) function that can be deployed readily in deep network architectures. Extensive experimental evaluations are conducted on three imbalanced person attribute benchmark datasets (CelebA, X-Domain, DeepFashion) and one balanced object category benchmark dataset (CIFAR-100). These experimental results demonstrate the performance advantages and model scalability of the proposed batch-wise incremental minority class rectification model over the existing state-of-the-art models for addressing the problem of imbalanced data learning.

Recognising detailed facial or clothing attributes in images of people is a challenging task for computer vision, especially when the training data are both in very large scale and extremely imbalanced among different attribute classes. To address this problem, we formulate a novel scheme for batch incremental hard sample mining of minority attribute classes from imbalanced large scale training data. We develop an end-to-end deep learning framework capable of avoiding the dominant effect of majority classes by discovering sparsely sampled boundaries of minority classes. This is made possible by introducing a Class Rectification Loss (CRL) regularising algorithm. We demonstrate the advantages and scalability of CRL over existing state-of-the-art attribute recognition and imbalanced data learning models on two large scale imbalanced benchmark datasets, the CelebA facial attribute dataset and the X-Domain clothing attribute dataset.

Magnetic reconnection is a process by which oppositely directed magnetic field lines passing through a plasma undergo dramatic rearrangement, converting magnetic potential into kinetic energy and heat 1,2. It is believed to play an important role in many plasma phenomena including solar flares 3,4 , star formation 5 and other astrophysical events 6 , laser-driven plasma jets 7-9 , and fusion plasma instabilities 10. Because of the large differences of scale between laboratory and astrophysical plasmas, it is often difficult to extrapolate the reconnection phenomena studied in one environment to those observed in the other. In some cases, however, scaling laws 11 do permit reliable connections to made, such as the experimental simulation of interactions between the solar wind and the Earth's magnetosphere 12. Here we report well-scaled laboratory experiments that reproduce loop-top-like X-ray source emission by reconnection outflows interacting with a solid target. Our experiments exploit the mega-gauss-scale magnetic field generated by interaction of a high-intensity laser with a plasma to reconstruct a magnetic reconnection topology similar to that which occurs in solar flares. We also identify the separatrix and diffusion regions associated with reconnection in which ions become decoupled from electrons on a scale of the ion inertial length. A major objective of laboratory astrophysics is to simulate the fundamental nature of astrophysical plasma physics processes in a laboratory environment so that certain astrophysical phenomenon can be studied in a controlled manner 13. High energy density facilities, such as high-powered lasers and Z-pinches, can provide such opportunities 14 , for example, direct measurements of opacity 15 , equations of state 16 , and photoionized plasmas 17,18 , as well as the similarity of physics, such as certain hydrodynamic phenomena of jets 19 and shocks 20 where a scaling law between astrophysical and laboratory plasma systems can be applied. As a fundamental cause of many plasma energy conversion processes, magnetic reconnection (MR) is certainly a high priority of such studies. Masuda et al. 21 observed the loop-top X-ray source in solar flares using the YOHKOH satellite and proposed that two antiparallel magnetic fields were merged above an arcade of closed loops as outflow jets from the reconnection point collided with high-density plasmas on the loop to produce a hot X-ray region. Ultraviolet 22 and X-ray 23,24 observations of plasma

Reconnection of the self-generated magnetic fields in laser-plasma interaction was first investigated experimentally by Nilson et al. [Phys. Rev. Lett. 97, 255001 (2006)] by shining two laser pulses a distance apart on a solid target layer. An elongated current sheet (CS) was observed in the plasma between the two laser spots. In order to more closely model magnetotail reconnection, here two side-by-side thin target layers, instead of a single one, are used. It is found that at one end of the elongated CS a fanlike electron outflow region including three well-collimated electron jets appears. The (>1 MeV) tail of the jet energy distribution exhibits a power-law scaling. The enhanced electron acceleration is attributed to the intense inductive electric field in the narrow electron dominated reconnection region, as well as additional acceleration as they are trapped inside the rapidly moving plasmoid formed in and ejected from the CS. The ejection also induces a secondary CS.