Search

In a spherically symmetric plasma constrained by its own gravity, the ionization degree lags behind changes in temperature and density. The ambipolar electric field accelerates ions radially and cools electrons. Ions lose energy and angular momentum in collisions with low-temperature electrons. The angular momentum of ions decreases much faster than their energy in cycles. The trajectories of ions are close to radial, and the density distribution resembles pseudo-isothermal. The velocity distributions of baryons moving outward and inward in galaxies tend to approach independent Maxwell distributions. This characteristic suppresses small-angle scattering of ions, reducing the ion collision cross-section by three orders of magnitude. In the massive galaxies, baryons can replace all dark matter., Comment: 10 pages, 4 figures

Cross-domain NER is a practical yet challenging problem since the data scarcity in the real-world scenario. A common practice is first to learn a NER model in a rich-resource general domain and then adapt the model to specific domains. Due to the mismatch problem between entity types across domains, the wide knowledge in the general domain can not effectively transfer to the target domain NER model. To this end, we model the label relationship as a probability distribution and construct label graphs in both source and target label spaces. To enhance the contextual representation with label structures, we fuse the label graph into the word embedding output by BERT. By representing label relationships as graphs, we formulate cross-domain NER as a graph matching problem. Furthermore, the proposed method has good applicability with pre-training methods and is potentially capable of other cross-domain prediction tasks. Empirical results on four datasets show that our method outperforms a series of transfer learning, multi-task learning, and few-shot learning methods., Comment: Findings of ACL; available at Findings 2022 https://aclanthology.org/2022.findings-acl.210/; Improve presentation

This paper introduces ConceptMath, a bilingual (English and Chinese), fine-grained benchmark that evaluates concept-wise mathematical reasoning of Large Language Models (LLMs). Unlike traditional benchmarks that evaluate general mathematical reasoning with an average accuracy, ConceptMath systematically organizes math problems under a hierarchy of math concepts, so that mathematical reasoning can be evaluated at different granularity with concept-wise accuracies. Based on our ConcepthMath, we evaluate a broad range of LLMs, and we observe existing LLMs, though achieving high average accuracies on traditional benchmarks, exhibit significant performance variations across different math concepts and may even fail catastrophically on the most basic ones. Besides, we also introduce an efficient fine-tuning strategy to enhance the weaknesses of existing LLMs. Finally, we hope ConceptMath could guide the developers to understand the fine-grained mathematical abilities of their models and facilitate the growth of foundation models., Comment: The benchmark dataset will be released soon

Why the temperature gradient in the vortex tube deviates significantly from the adiabatic gradient is a very important but unresolved issue in vortex tube research. In compressible fluids, the vorticity of a fluid parcel that is different from its environment can affect its own pressure or density, just like temperature. Gases near the wall have very high vorticity, and when they detach from the wall, their higher vorticity compared to the environment makes their density lower than the environment. Driven by vorticity, some fluid parcels undergo multiple collisions near the axis, achieving temperature separation., Comment: 6pages, 1figre

Fine-tuning has been proven to be a simple and effective technique to transfer the learned knowledge of Pre-trained Language Models (PLMs) to downstream tasks. However, vanilla fine-tuning easily overfits the target data and degrades the generalization ability. Most existing studies attribute it to catastrophic forgetting, and they retain the pre-trained knowledge indiscriminately without identifying what knowledge is transferable. Motivated by this, we frame fine-tuning into a causal graph and discover that the crux of catastrophic forgetting lies in the missing causal effects from the pretrained data. Based on the causal view, we propose a unified objective for fine-tuning to retrieve the causality back. Intriguingly, the unified objective can be seen as the sum of the vanilla fine-tuning objective, which learns new knowledge from target data, and the causal objective, which preserves old knowledge from PLMs. Therefore, our method is flexible and can mitigate negative transfer while preserving knowledge. Since endowing models with commonsense is a long-standing challenge, we implement our method on commonsense QA with a proposed heuristic estimation to verify its effectiveness. In the experiments, our method outperforms state-of-the-art fine-tuning methods on all six commonsense QA datasets and can be implemented as a plug-in module to inflate the performance of existing QA models., Comment: ACL 2023 (oral paper)

Based on the rotating turbulent thermal convection model and using the rotating equivalent temperature assumption and new convection criterion, this paper analyzed the repression of the vorticity gradient on the heat transport and explained that the formation of sunspots originated from the variation of the vorticity gradient in the solar troposphere., Comment: 8 pages

The rotation of granules shows thermal properties similar to the thermal motion of molecules, which effects convection criterion and solar granule size.The granular rotation generates the rotational additional pressure and corresponding rotational equivalent temperature in the spherically symmetric expansion of the granule.The rotation equivalent temperature represented by the rotational speed and the granular size is combined with the Schwarzschild criterion to obtain a new convection criterion determined jointly by the granular size, temperature gradient, and rotational speed gradient.In most of the solar convection zone, the convection is driven by the temperature gradient and suppressed by the rotational speed gradient.In the solar quiet region, the larger the granular size, the weaker the convection.The small granules with the sizes smaller than the critical size are in natural convection, while the large granules with the sizes larger than the critical size are in forced convection.In the solar active region, the larger the rotational speed, the smaller the critical size and the temperature.So sunspots have smaller granules and lower temperatures., Comment: 7 pages, 5 figures, Submitted to MNRAS

Continual Learning for Named Entity Recognition (CL-NER) aims to learn a growing number of entity types over time from a stream of data. However, simply learning Other-Class in the same way as new entity types amplifies the catastrophic forgetting and leads to a substantial performance drop. The main cause behind this is that Other-Class samples usually contain old entity types, and the old knowledge in these Other-Class samples is not preserved properly. Thanks to the causal inference, we identify that the forgetting is caused by the missing causal effect from the old data. To this end, we propose a unified causal framework to retrieve the causality from both new entity types and Other-Class. Furthermore, we apply curriculum learning to mitigate the impact of label noise and introduce a self-adaptive weight for balancing the causal effects between new entity types and Other-Class. Experimental results on three benchmark datasets show that our method outperforms the state-of-the-art method by a large margin. Moreover, our method can be combined with the existing state-of-the-art methods to improve the performance in CL-NER, Comment: Accepted by EMNLP2022

The expansion of the rotating fluid will change the vorticity and rotational speed of the expanding region. In turbulent thermal convection, this microscopic effect is preserved. Tracking the fluid micelles shows that the average vorticity varies with density, producing vorticity transport and angular momentum transport from the low-density area to the high-density area, forming a macroscopic vorticity difference and rotational speed difference. Taking the axial thermal convection model of the solar polar region, it can generate axial differential rotation, and the centrifugal force difference generated by the axial differential rotation drives the meridional circulation, transporting angular momentum away from the axis of rotation, forming latitudinal differential rotation. The rotation of the fluid cell generates additional pressure and change the convection criterion. The temperature gradient in the solar troposphere is higher than in the non-rotating fluid model, It's the energy source of the differential rotating.