Search

Large Language Models (LLMs) have shown remarkable capabilities in various natural language processing tasks. However, LLMs may rely on dataset biases as shortcuts for prediction, which can significantly impair their robustness and generalization capabilities. This paper presents Shortcut Suite, a comprehensive test suite designed to evaluate the impact of shortcuts on LLMs' performance, incorporating six shortcut types, five evaluation metrics, and four prompting strategies. Our extensive experiments yield several key findings: 1) LLMs demonstrate varying reliance on shortcuts for downstream tasks, significantly impairing their performance. 2) Larger LLMs are more likely to utilize shortcuts under zero-shot and few-shot in-context learning prompts. 3) Chain-of-thought prompting notably reduces shortcut reliance and outperforms other prompting strategies, while few-shot prompts generally underperform compared to zero-shot prompts. 4) LLMs often exhibit overconfidence in their predictions, especially when dealing with datasets that contain shortcuts. 5) LLMs generally have a lower explanation quality in shortcut-laden datasets, with errors falling into three types: distraction, disguised comprehension, and logical fallacy. Our findings offer new insights for evaluating robustness and generalization in LLMs and suggest potential directions for mitigating the reliance on shortcuts. The code is available at \url {https://github.com/yyhappier/ShortcutSuite.git}.

The COVID-19 pandemic highlighted the urgent need for effective vaccine policies, but traditional clinical trials often lack sufficient data to capture the diverse population characteristics necessary for comprehensive public health strategies. Ethical concerns around randomized trials during a pandemic further complicate policy development for public health. Reinforcement Learning (RL) offers a promising alternative for vaccine policy development. However, direct online RL exploration in real-world scenarios can result in suboptimal and potentially harmful decisions. This study proposes a novel framework combining tabular Q-learning with microsimulation (i.e., a Recurrent Neural Network (RNN) environment simulator) to address these challenges in public health vaccine policymaking, which enables effective vaccine policy learning without real-world interaction, addressing both ethical and exploration challenges. The RNN environment simulator captures temporal associations between infection and patient characteristics, generating realistic simulation data. Our tabular Q-learning model produces an interpretable policy table that balances the risks of severe infection against vaccination side effects. Applied to COVID-19 booster policies, the learned Q-learning-based policy outperforms current practices, offering a path toward more effective vaccination strategies.

Interlayer coupling in two-dimensional (2D) layered nanomaterials can provide us novel strategies to evoke their superior properties, such as the exotic flat bands and unconventional superconductivity of twisted layers, the formation of moir\'e excitons and related nontrivial topology. However, to accurately quantify interlayer potential and further measure elastic properties of 2D materials remains vague, despite significant efforts. Herein, the layer-dependent lattice dynamics and elastic constants of 2D nanomaterials have been systematically investigated via pressure-engineering strategy based on ultralow frequency Raman spectroscopy. The shearing mode and layer-breathing Raman shifts of MoS2 with various thicknesses were analyzed by the linear chain model. Intriguingly, it was found that the layer-dependent d{\omega}/dP of shearing and breathing Raman modes display the opposite trends, quantitatively consistent with our molecular dynamics simulations and density functional theory calculations. These results can be generalized to other van der Waals systems, and may shed light on the potential applications of 2D materials in nanomechanics and nanoelectronics., Comment: 25 pages, 5 figures

With the development of legal intelligence, Criminal Court View Generation has attracted much attention as a crucial task of legal intelligence, which aims to generate concise and coherent texts that summarize case facts and provide explanations for verdicts. Existing researches explore the key information in case facts to yield the court views. Most of them employ a coarse-grained approach that partitions the facts into broad segments (e.g., verdict-related sentences) to make predictions. However, this approach fails to capture the complex details present in the case facts, such as various criminal elements and legal events. To this end, in this paper, we propose an Event Grounded Generation (EGG) method for criminal court view generation with cooperative (Large) Language Models, which introduces the fine-grained event information into the generation. Specifically, we first design a LLMs-based extraction method that can extract events in case facts without massive annotated events. Then, we incorporate the extracted events into court view generation by merging case facts and events. Besides, considering the computational burden posed by the use of LLMs in the extraction phase of EGG, we propose a LLMs-free EGG method that can eliminate the requirement for event extraction using LLMs in the inference phase. Extensive experimental results on a real-world dataset clearly validate the effectiveness of our proposed method., Comment: Accepted to SIGIR2024

We study two dimensional electron systems confined in wide quantum wells whose subband separation is comparable with the Zeeman energy. Two N = 0 Landau levels from different subbands and with opposite spins are pinned in energy when they cross each other and electrons can freely transfer between them. When the disorder is strong, we observe clear hysteresis in our data corresponding to instability of the electron distribution in the two crossing levels. When the intra-layer interaction dominates, multiple minima appear when a Landau level is 1/3 or 2/3 filled and fractional quantum hall effect can be stabilized.

In this paper, we study Kropina metrics with isotropic scalar curvature. First, we obtain the expressions of Ricci curvature tensor and scalar curvature. Then, we characterize the Kropina metrics with isotropic scalar curvature on by tensor analysis.

Objective: Bleeding from gastroesophageal varices (GEV) is a medical emergency associated with high mortality. We aim to construct an artificial intelligence-based model of two-dimensional shear wave elastography (2D-SWE) of the liver and spleen to precisely assess the risk of GEV and high-risk gastroesophageal varices (HRV). Design: A prospective multicenter study was conducted in patients with compensated advanced chronic liver disease. 305 patients were enrolled from 12 hospitals, and finally 265 patients were included, with 1136 liver stiffness measurement (LSM) images and 1042 spleen stiffness measurement (SSM) images generated by 2D-SWE. We leveraged deep learning methods to uncover associations between image features and patient risk, and thus conducted models to predict GEV and HRV. Results: A multi-modality Deep Learning Risk Prediction model (DLRP) was constructed to assess GEV and HRV, based on LSM and SSM images, and clinical information. Validation analysis revealed that the AUCs of DLRP were 0.91 for GEV (95% CI 0.90 to 0.93, p < 0.05) and 0.88 for HRV (95% CI 0.86 to 0.89, p < 0.01), which were significantly and robustly better than canonical risk indicators, including the value of LSM and SSM. Moreover, DLPR was better than the model using individual parameters, including LSM and SSM images. In HRV prediction, the 2D-SWE images of SSM outperform LSM (p < 0.01). Conclusion: DLRP shows excellent performance in predicting GEV and HRV over canonical risk indicators LSM and SSM. Additionally, the 2D-SWE images of SSM provided more information for better accuracy in predicting HRV than the LSM.

The vaccine adverse event reporting system (VAERS) is a vital resource for post-licensure vaccine safety monitoring and has played a key role in assessing the safety of COVID-19 vaccines. However it is difficult to properly identify rare adverse events (AEs) associated with vaccines due to small or zero counts. We propose a Bayesian model with a Dirichlet Process Mixture prior to improve accuracy of the AE estimates with small counts by allowing data-guided information sharing between AE estimates. We also propose a negative control procedure embedded in our Bayesian model to mitigate the reporting bias due to the heightened awareness of COVID-19 vaccines, and use it to identify associated AEs as well as associated AE groups defined by the organ system in the Medical Dictionary for Regulatory Activities (MedDRA) ontology. The proposed model is evaluated using simulation studies, in which it outperforms baseline models without information sharing and is applied to study the safety of COVID-19 vaccines using VAERS data.

The Wigner crystal, an ordered array of electrons, is one of the very first proposed many-body phases stabilized by the electron-electron interaction. This electron solid phase has been reported in ultra-clean two-dimensional electron systems at extremely low temperatures, where the Coulomb interaction dominants over the kinetic energy, disorder potential and thermal fluctuation. We closely examine this quantum phase with capacitance measurements where the device length-scale is comparable with the crystal's correlation length. The extraordinarily high performance of our technique makes it possible to quantitatively study the dynamic response of the Wigner crystal within the single crystal regime. Our result will greatly boost the study of this inscrutable electron solid.

While vaccines are crucial to end the COVID-19 pandemic, public confidence in vaccine safety has always been vulnerable. Many statistical methods have been applied to VAERS (Vaccine Adverse Event Reporting System) database to study the safety of COVID-19 vaccines. However, all these methods ignored the adverse event (AE) ontology. AEs are naturally related; for example, events of retching, dysphagia, and reflux are all related to an abnormal digestive system. Explicitly bringing AE relationships into the model can aid in the detection of true AE signals amid the noise while reducing false positives. We propose a Bayesian graphical model to estimate all AEs while incorporating the AE ontology simultaneously. We proposed strategies to construct conjugate forms leading to an efficient Gibbs sampler. Built upon the posterior distributions, we proposed a negative control approach to mitigate reporting bias and an enrichment approach to detect AE groups of concern. The proposed methods were evaluated using simulation studies and were further illustrated on studying the safety of COVID-19 vaccines. The proposed methods were implemented in R package \textit{BGrass} and source code are available at https://github.com/BangyaoZhao/BGrass., Comment: 12 pages, 5 figures

Capacitance measurement is a useful technique in studying quantum devices, as it directly probes the local particle charging properties, i.e. the system compressibility. Here we report one approach which can measure capacitance from mK to room temperature with excellent accuracy. Our experiments show that such a high-precision technique is able to reveal delicate and essential properties of high-mobility two-dimensional electron systems.

In this report, we study ultra-high-mobility two-dimensional (2D) electron/hole systems with high precision capacitance measurement. We find that the capacitance charge appears only at the fringe of the gate at high magnetic field when the 2D conductivity decreases significantly. At integer quantum Hall effects, the capacitance vanishes and forms a plateau at high temperatures $T\gtrsim 300$ mK, which surprisingly, disappears at $T\lesssim 100$ mK. This anomalous behavior is likely a manifestation that the dilute particles/vacancies in the top-most Landau level form Wigner crystal, which has finite compressibility and can host polarization current.

Modern video-text retrieval frameworks basically consist of three parts: video encoder, text encoder and the similarity head. With the success on both visual and textual representation learning, transformer based encoders and fusion methods have also been adopted in the field of video-text retrieval. In this report, we present CLIP2TV, aiming at exploring where the critical elements lie in transformer based methods. To achieve this, We first revisit some recent works on multi-modal learning, then introduce some techniques into video-text retrieval, finally evaluate them through extensive experiments in different configurations. Notably, CLIP2TV achieves 52.9@R1 on MSR-VTT dataset, outperforming the previous SOTA result by 4.1%.