Search

ChatGPT, a language-learning model chatbot, has garnered considerable attention for its ability to respond to users’ questions. Using data from 14 countries and 186 institutions, we compare ChatGPT and student performance for 28,085 questions from accounting assessments and textbook test banks. As of January 2023, ChatGPT provides correct answers for 56.5 percent of questions and partially correct answers for an additional 9.4 percent of questions. When considering point values for questions, students significantly outperform ChatGPT with a 76.7 percent average on assessments compared to 47.5 percent for ChatGPT if no partial credit is awarded and 56.5 percent if partial credit is awarded. Still, ChatGPT performs better than the student average for 15.8 percent of assessments when we include partial credit. We provide evidence of how ChatGPT performs on different question types, accounting topics, class levels, open/closed assessments, and test bank questions. We also discuss implications for accounting education and research.

CDEX-50 is a next-generation project of the China Dark Matter Experiment (CDEX) that aims to search for dark matter using a 50-kg germanium detector array. This paper comprises a thorough summary of the CDEX-50 experiment, including an investigation of potential background sources and the development of a background model. Based on the baseline model, the projected sensitivity of weakly interacting massive particle (WIMP) is also presented. The expected background level within the energy region of interest, set to 2--2.5 keVee, is $\sim$0.01 counts keVee$^{-1}$ kg$^{-1}$ day$^{-1}$. At 90\% confidence level, the expected sensitivity to spin-independent WIMP-nucleon couplings is estimated to reach a cross-section of 5.3 $\times$ 10$^{-45}$ cm$^{2}$ for a WIMP mass of 5 GeV/c$^{2}$ with an exposure objective of 150 kg$\cdot$year and an analysis threshold of 160 eVee. This science goal will correspond to the most sensitive results for WIMPs with a mass of 2.2--8 GeV/c$^{2}$., Comment: 11 pages, 11 figures

A neural ranker plays an indispensable role in the de facto 'retrieval & rerank' pipeline, but its training still lags behind due to the weak negative mining during contrastive learning. Compared to retrievers boosted by self-adversarial (i.e., in-distribution) negative mining, the ranker's heavy structure suffers from query-document combinatorial explosions, so it can only resort to the negative sampled by the fast yet out-of-distribution retriever. Thereby, the moderate negatives compose ineffective contrastive learning samples, becoming the main barrier to learning a robust ranker. To alleviate this, we propose a multi-adversarial training strategy that leverages multiple retrievers as generators to challenge a ranker, where i) diverse hard negatives from a joint distribution are prone to fool the ranker for more effective adversarial learning and ii) involving extensive out-of-distribution label noises renders the ranker against each noise distribution, leading to more challenging and robust contrastive learning. To evaluate our robust ranker (dubbed R2ANKER), we conduct experiments in various settings on the passage retrieval benchmarks, including BM25-reranking, full-ranking, retriever distillation, etc. The empirical results verify the new state-of-the-art effectiveness of our model.

Large-scale retrieval is to recall relevant documents from a huge collection given a query. It relies on representation learning to embed documents and queries into a common semantic encoding space. According to the encoding space, recent retrieval methods based on pre-trained language models (PLM) can be coarsely categorized into either dense-vector or lexicon-based paradigms. These two paradigms unveil the PLMs' representation capability in different granularities, i.e., global sequence-level compression and local word-level contexts, respectively. Inspired by their complementary global-local contextualization and distinct representing views, we propose a new learning framework, Unifier, which unifies dense-vector and lexicon-based retrieval in one model with a dual-representing capability. Experiments on passage retrieval benchmarks verify its effectiveness in both paradigms. A uni-retrieval scheme is further presented with even better retrieval quality. We lastly evaluate the model on BEIR benchmark to verify its transferability.

We present new constraints on light dark matter boosted by cosmic rays (CRs) using the 205.4 kg day data of the CDEX-10 experiment conducted at the China Jinping Underground Laboratory. The Monte-Carlo simulation package CJPL\_ESS was employed to evaluate the Earth shielding effect. Several key factors have been introduced and discussed in our CRDM analysis, including the contributions from heavier CR nuclei than proton and helium, the inhomogeneity of CR distribution, and the impact of the form factor in the Earth attenuation calculation. Our result excludes the dark matter--nucleon elastic scattering cross-section region from $1.7\times 10^{-30}~\rm cm^2$ to $10^{-26}~\rm cm^2$ for dark matter of 10 keV$/c^2$ to 1 GeV$/c^2$., Comment: 9 pages, 7 figures. Added discussion of CR inputs and parameter choices

This work studies temporal reading comprehension (TRC), which reads a free-text passage and answers temporal ordering questions. Precise question understanding is critical for temporal reading comprehension. For example, the question “What happened before the victory” and “What happened after the victory” share almost all words except one, while their answers are totally different. Moreover, even if two questions query about similar temporal relations, different varieties might also lead to various answers. For example, although both the question “What usually happened during the press release?” and “What might happen during the press release” query events which happen after the press release, they convey divergent semantics. To this end, we propose a novel reading comprehension approach with precise question understanding. Specifically, a temporal ordering question is embedded into two vectors to capture the referred event and the temporal relation. Then we evaluate the temporal relation between candidate events and the referred event based on that. Such fine-grained representations offer two benefits. First, it enables a better understanding of the question by focusing on different elements of a question. Second, it provides good interpretability when evaluating temporal relations. Furthermore, we also harness an auxiliary contrastive loss for representation learning of temporal relations, which aims to distinguish relations with subtle but critical changes. The proposed approach outperforms strong baselines and achieves state-of-the-art performance on the TORQUE dataset. It also increases the accuracy of four pre-trained language models (BERT base, BERT large, RoBERTa base, and RoBETRa large), demonstrating its generic effectiveness on divergent models.

When demonstrating the effectiveness of a new algorithm, researchers are traditionally encouraged to compare their algorithm's performance against existing algorithms on well-studied benchmark test suites. In the absence of more nuanced methodologies, algorithm performance is typically summarized on average across the test suite examples. This paper highlights the potential bias of conclusions drawn by analyzing "on average" performance, and the opportunities offered by a recent testing methodology known as instance space analysis. To illustrate, we revisit our 2007 comparative study of algorithms for facial age estimation, and rigorously stress-test to challenge the original conclusions. The case study demonstrates how powerful visualizations offered by instance space analysis enable greater insights into unique strengths and weaknesses, and which algorithm should be used when and why. Inspired by such insights, a new algorithm is proposed, and its unique advantage is demonstrated. The bias often hidden in well-studied datasets, and the ramifications for drawing biased conclusions, are also illustrated in this case study. While focused on facial age estimation, the methodology and lessons learned from the case study are broadly applicable to any study seeking to draw conclusions about algorithm performance based on empirical results.

Generating new events given context with correlated ones plays a crucial role in many event-centric reasoning tasks. Existing works either limit their scope to specific scenarios or overlook event-level correlations. In this paper, we propose to pre-train a general Correlation-aware context-to-Event Transformer (ClarET) for event-centric reasoning. To achieve this, we propose three novel event-centric objectives, i.e., whole event recovering, contrastive event-correlation encoding and prompt-based event locating, which highlight event-level correlations with effective training. The proposed ClarET is applicable to a wide range of event-centric reasoning scenarios, considering its versatility of (i) event-correlation types (e.g., causal, temporal, contrast), (ii) application formulations (i.e., generation and classification), and (iii) reasoning types (e.g., abductive, counterfactual and ending reasoning). Empirical fine-tuning results, as well as zero- and few-shot learning, on 9 benchmarks (5 generation and 4 classification tasks covering 4 reasoning types with diverse event correlations), verify its effectiveness and generalization ability.

Deep neural network-based models have achieved great success in extractive question answering. Recently, many works have been proposed to model multistage matching for this task, which usually first retrieve relevant paragraphs or sentences and then extract an answer span from the retrieved results. However, such a pipeline-based approach suffers from the error propagation problem, especially for sentence-level retrieval that is usually difficult to achieve high accuracy due to the severe data imbalance problem. Furthermore, since the paragraph/sentence selector and the answer extractor are closely related, modeling them independently does not fully exploit the power of multistage matching. To solve these problems, we propose a novel end-to-end multigranularity reading comprehension model, which is a unified framework to explicitly model three matching granularities, including paragraph identification, sentence selection, and answer extraction. Our approach has two main advantages. First, the end-to-end approach alleviates the error propagation problem in both the training and inference phases. Second, the shared features in a unified model improve the learning of representations of different matching granularities. We conduct a comprehensive comparison on four large-scale datasets (SQuAD-open, NewsQA, SQuAD 2.0, and SQuAD Adversarial) and verify that the proposed approach outperforms both the vanilla BERT model and existing multistage matching approaches. We also conduct an ablation study and verify the effectiveness of the proposed components in our model structure.

Event correlation reasoning infers whether a natural language paragraph containing multiple events conforms to human common sense. For example, "Andrew was very drowsy, so he took a long nap, and now he is very alert"is sound and reasonable. In contrast, "Andrew was very drowsy, so he stayed up a long time, now he is very alert"does not comply with human common sense. Such reasoning capability is essential for many downstream tasks, such as script reasoning, abductive reasoning, narrative incoherence, story cloze test, etc. However, conducting event correlation reasoning is challenging due to a lack of large amounts of diverse event-based knowledge and difficulty in capturing correlation among multiple events. In this paper, we propose EventBERT, a pre-trained model to encapsulate eventuality knowledge from unlabeled text. Specifically, we collect a large volume of training examples by identifying natural language paragraphs that describe multiple correlated events and further extracting event spans in an unsupervised manner. We then propose three novel event- and correlation-based learning objectives to pre-train an event correlation model on our created training corpus. Experimental results show EventBERT outperforms strong baselines on four downstream tasks, and achieves state-of-the-art results on most of them. Moreover, it outperforms existing pre-trained models by a large margin, e.g., 6.5 23%, in zero-shot learning of these tasks.