Multiple entities in a document generally exhibit complex inter-sentence relations, and cannot be well handled by existing relation extraction (RE) methods that typically focus on extracting intra-sentence relations for single entity pairs. In order to accelerate the research on document-level RE, we introduce DocRED, a new dataset constructed from Wikipedia and Wikidata with three features: (1) DocRED annotates both named entities and relations, and is the largest humanannotated dataset for document-level RE from plain text; (2) DocRED requires reading multiple sentences in a document to extract entities and infer their relations by synthesizing all information of the document; (3) along with the human-annotated data, we also offer large-scale distantly supervised data, which enables DocRED to be adopted for both supervised and weakly supervised scenarios. In order to verify the challenges of documentlevel RE, we implement recent state-of-the-art methods for RE and conduct a thorough evaluation of these methods on DocRED. Empirical results show that DocRED is challenging for existing RE methods, which indicates that document-level RE remains an open problem and requires further efforts. Based on the detailed analysis on the experiments, we discuss multiple promising directions for future research. We make DocRED and the code for our baselines publicly available at https: //github.com/thunlp/DocRED. * indicates equal contribution

Low-cost and high-efficiency solar cells are attractive candidates to meet the growing demand for renewable energy. Silicon solar cells have reached a power conversion efficiency (PCE) of 26.6%, [1] approaching their theoretical limit; tandem solar cells offer a means to improve further the efficiency of solar cells.Large-bandgap perovskites offer a route to improve the efficiency of energy capture in photovoltaics when employed in the front cell of perovskite-silicon tandems. Implementing perovskites as the front cell requires an inverted (p-i-n) architecture; this architecture is particularly effective at harnessing highenergy photons and is compatible with ionic-dopant-free transport layers. Here, a power conversion efficiency of 21.6% is reported, the highest among inverted perovskite solar cells (PSCs). Only by introducing a secondary amine into the perovskite structure to form MA 1−x DMA x PbI 3 (MA is methylamine and DMA is dimethylamine) are defect density and carrier recombination suppressed to enable record performance. It is also found that the controlled inclusion of DMA increases the hydrophobicity and stability of films in ambient operating conditions: encapsulated devices maintain over 80% of their efficiency following 800 h of operation at the maximum power point, 30 times longer than reported in the best prior inverted PSCs. The unencapsulated devices show record operational stability in ambient air among PSCs.

No abstract

Fact Verification requires fine-grained natural language inference capability that finds subtle clues to identify the syntactical and semantically correct but not well-supported claims. This paper presents Kernel Graph Attention Network (KGAT), which conducts more finegrained fact verification with kernel-based attentions. Given a claim and a set of potential evidence sentences that form an evidence graph, KGAT introduces node kernels, which better measure the importance of the evidence node, and edge kernels, which conduct fine-grained evidence propagation in the graph, into Graph Attention Networks for more accurate fact verification. KGAT achieves a 70.38% FEVER score and significantly outperforms existing fact verification models on FEVER, a large-scale benchmark for fact verification. Our analyses illustrate that, compared to dot-product attentions, the kernelbased attention concentrates more on relevant evidence sentences and meaningful clues in the evidence graph, which is the main source of KGAT's effectiveness. All source codes of this work are available at https://github. com/thunlp/KernelGAT. Al Jardine is an American rhythm guitarist Claim Verification SUPPORTS REFUTES NOT ENOUGH INFO Evidence Reasoning Alan Charles Jardine (born September 3, 1942) is an American musician, singer and songwriter who cofounded the Beach Boys. He is best known as the band's rhythm guitarist, and for occasionally singing lead vocals on singles.

Solid-state color centers with manipulable spin qubits and telecom-ranged fluorescence are ideal platforms for quantum communications and distributed quantum computations. In this work, we coherently control the nitrogen-vacancy (NV) center spins in silicon carbide at room temperature, in which the telecomwavelength emission is detected. Through carefully optimizing the implanted conditions, we improve the concentration of NV centers for about 4 times. Based on this, the coherent control of NV center spins is achieved at room temperature and the coherence time T2 * can be reached around 1 μs. Furthermore, the investigation of fluorescence properties of single NV centers shows that they are room temperature photostable single photon sources at telecom range. Taking the advantages of the technological mature materials, the experiment demonstrates that the NV centers in silicon carbide are promising systems for large-scale integrated quantum photonics and long-distance quantum networks.

Language representation models such as BERT could effectively capture contextual semantic information from plain text, and have been proved to achieve promising results in lots of downstream NLP tasks with appropriate fine-tuning. However, most existing language representation models cannot explicitly handle coreference, which is essential to the coherent understanding of the whole discourse. To address this issue, we present CorefBERT, a novel language representation model that can capture the coreferential relations in context. The experimental results show that, compared with existing baseline models, CorefBERT can achieve significant improvements consistently on various downstream NLP tasks that require coreferential reasoning, while maintaining comparable performance to previous models on other common NLP tasks. The source code and experiment details of this paper can be obtained from https://github. com/thunlp/CorefBERT.

Human conversations naturally evolve around related concepts and scatter to multi-hop concepts. This paper presents a new conversation generation model, ConceptFlow, which leverages commonsense knowledge graphs to explicitly model conversation flows. By grounding conversations to the concept space, Con-ceptFlow represents the potential conversation flow as traverses in the concept space along commonsense relations. The traverse is guided by graph attentions in the concept graph, moving towards more meaningful directions in the concept space, in order to generate more semantic and informative responses. Experiments on Reddit conversations demonstrate ConceptFlow's effectiveness over previous knowledge-aware conversation models and GPT-2 based models while using 70% fewer parameters, confirming the advantage of explicit modeling conversation structures. All source codes of this work are available at https://github.com/ thunlp/ConceptFlow.

The hole transport material (HTM) free carbon based perovskite solar cells (C-PSCs) are promising for its manufactural simplicity, but they currently suffer from low power conversion efficiencies (PCE) largely because of the voltage loss. Here, a new strategy to increase the PCE by incorporating an ultrathin ferroelectric oxide PbTiO 3 layer between the electron transport material and the halide perovskite is reported. The resulting C-PSCs have achieved PCEs up to 16.37%, which is the highest record for HTM-free C-PSCs to date, mainly ascribable to the ferroelectric layer enhanced open circuit voltage. Detail measurements and analysis show an enhanced built-in potential in the C-PSCs as well as suppression of the non-radiative recombination due to the ferroelectric PbTiO 3 layer incorporation, accounting for the boosted V OC and photovoltaic performance.