Detecting negatives (such as non-entailment relationships, unanswerable questions, and false claims) is an important and challenging aspect of many natural language understanding tasks. Though manually collecting challenging negative examples can help models detect them, it is both costly and domain-specific. In this work, we propose Self-labeled Counterfactuals for Extrapolating to Negative Examples (SCENE), an automatic method for synthesizing training data that greatly improves models' ability to detect challenging negative examples. In contrast with standard data augmentation, which synthesizes new examples for existing labels, SCENE can synthesize negative examples zero-shot from only positive ones. Given a positive example, SCENE perturbs it with a mask infilling model, then determines whether the resulting example is negative based on a self-training heuristic. With access to only answerable training examples, SCENE can close 69.6% of the performance gap on SQuAD 2.0, a dataset where half of the evaluation examples are unanswerable, compared to a model trained on SQuAD 2.0. Our method also extends to boolean question answering and recognizing textual entailment, and improves generalization from SQuAD to ACE-whQA, an out-of-domain extractive QA benchmark.

Ukraine's air force said Russia launched 14 attack drones and five missiles on the southern Black Sea regions with air defence systems destroying 11 of the drones. The Ministry of Internal Affairs of Ukraine said six people were injured in the historic city of Odesa and residential buildings and a warehouse were damaged. Ukrainian security sources said they orchestrated a drone attack on an oil refinery in the southern Russian town of Tuapse, about 240 kilometres (150 miles) southeast of the Russian-annexed Crimean peninsula. The attack caused a major fire, but there were no reports of casualties. Nepal's Foreign Minister Narayan Prakash Saud told the Associated Press news agency that Nepal had asked Russia to send back hundreds of Nepali nationals who had been recruited to fight against Ukraine and repatriate the bodies of those who had died in the conflict.

We introduce the spherical convolutional historical simulations coincide with observational measurements, Wasserstein distance to more comprehensively we can compare each model's synthetic climate measure differences between climate models and distribution to the distribution of observational or quasiobservational reanalysis data. This new similarity measure accounts data products (Raäisaänen, 2007), to assess for spatial variability using convolutional their reconstructive skill. For complete spatial coverage we projections and quantifies local differences in the compare against reanalysis data, a blend of observations distribution of climate variables. We apply this and short-range weather forecasts through data assimilation method to evaluate the historical model outputs (Bengtsson et al., 2004). This has become one popular of the Coupled Model Intercomparison Project climate model validation method (Flato et al., 2014).

Recommendation 1: Develop Hybrid AI-Physical Models: Emphasize the integration of AI and physical modeling for improved reliability, especially for longer prediction horizons, acknowledging the delicate balance between knowledge-based and data-driven components required for optimal performance. Recommendation 2: Emphasize Robustness in AI Downscaling Approaches, favoring techniques that respect physical laws, preserve inter-variable dependencies and spatial structures, and accurately represent extremes at the local scale. Recommendation 3: Promote Inclusive Model Development: Ensure Earth System Model development is open and accessible to diverse stakeholders, enabling forecasters, the public, and AI/statistics experts to use, develop, and engage with the model and its predictions/projections. Figure Caption: Advancements in data collection, data access, hybrid AI-physical Earth system modeling, and downscaling empower stakeholders with increased accessibility to local predictions and projections, encouraging collaborative efforts across disciplines to improve climate change preparedness. Here, we review how machine learning has interactions (Rosenfeld et al., 2014). In the ocean, uncertainties persist due that can be integrated forward in time, serve the to unresolved mesoscale eddies and turbulent double purpose of understanding and prediction processes (Couldrey et al., 2021).

WeatherBench 2 is an update to the global, medium-range (1-14 day) weather forecasting benchmark proposed by Rasp et al. (2020), designed with the aim to accelerate progress in data-driven weather modeling. WeatherBench 2 consists of an open-source evaluation framework, publicly available training, ground truth and baseline data as well as a continuously updated website with the latest metrics and state-of-the-art models: https://sites.research.google/weatherbench. This paper describes the design principles of the evaluation framework and presents results for current state-of-the-art physical and data-driven weather models. The metrics are based on established practices for evaluating weather forecasts at leading operational weather centers. We define a set of headline scores to provide an overview of model performance. In addition, we also discuss caveats in the current evaluation setup and challenges for the future of data-driven weather forecasting.

Modern plankton high-throughput monitoring relies on deep learning classifiers for species recognition in water ecosystems. Despite satisfactory nominal performances, a significant challenge arises from Dataset Shift, which causes performances to drop during deployment. In our study, we integrate the ZooLake dataset with manually-annotated images from 10 independent days of deployment, serving as test cells to benchmark Out-Of-Dataset (OOD) performances. Our analysis reveals instances where classifiers, initially performing well in In-Dataset conditions, encounter notable failures in practical scenarios. For example, a MobileNet with a 92% nominal test accuracy shows a 77% OOD accuracy. We systematically investigate conditions leading to OOD performance drops and propose a preemptive assessment method to identify potential pitfalls when classifying new data, and pinpoint features in OOD images that adversely impact classification. We present a three-step pipeline: (i) identifying OOD degradation compared to nominal test performance, (ii) conducting a diagnostic analysis of degradation causes, and (iii) providing solutions. We find that ensembles of BEiT vision transformers, with targeted augmentations addressing OOD robustness, geometric ensembling, and rotation-based test-time augmentation, constitute the most robust model, which we call BEsT model. It achieves an 83% OOD accuracy, with errors concentrated on container classes. Moreover, it exhibits lower sensitivity to dataset shift, and reproduces well the plankton abundances. Our proposed pipeline is applicable to generic plankton classifiers, contingent on the availability of suitable test cells. By identifying critical shortcomings and offering practical procedures to fortify models against dataset shift, our study contributes to the development of more reliable plankton classification technologies.

Large language models (LLMs) often generate convincing, fluent explanations. However, different from humans, they often generate inconsistent explanations on different inputs. For example, an LLM may generate the explanation "all birds can fly" when answering the question "Can sparrows fly?" but meanwhile answer "no" to the related question "Can penguins fly?". Explanations should be consistent across related examples so that they allow a human to simulate the LLM's decision process on multiple examples. We propose explanation-consistency finetuning (EC-finetuning), a method that adapts LLMs to generate more consistent natural-language explanations on related examples. EC-finetuning involves finetuning LLMs on synthetic data that is carefully constructed to contain consistent explanations. Across a variety of question-answering datasets in various domains, EC-finetuning yields a 10.0% relative explanation consistency improvement on four finetuning datasets, and generalizes to seven out-of-distribution datasets not seen during finetuning (+4.5% relative). Code is available at https://github.com/yandachen/explanation-consistency-finetuning .

Large language models (LLMs), exemplified by ChatGPT, have gained considerable attention for their excellent natural language processing capabilities. Nonetheless, these LLMs present many challenges, particularly in the realm of trustworthiness. Therefore, ensuring the trustworthiness of LLMs emerges as an important topic. This paper introduces TrustLLM, a comprehensive study of trustworthiness in LLMs, including principles for different dimensions of trustworthiness, established benchmark, evaluation, and analysis of trustworthiness for mainstream LLMs, and discussion of open challenges and future directions. Specifically, we first propose a set of principles for trustworthy LLMs that span eight different dimensions. Based on these principles, we further establish a benchmark across six dimensions including truthfulness, safety, fairness, robustness, privacy, and machine ethics. We then present a study evaluating 16 mainstream LLMs in TrustLLM, consisting of over 30 datasets. Our findings firstly show that in general trustworthiness and utility (i.e., functional effectiveness) are positively related. Secondly, our observations reveal that proprietary LLMs generally outperform most open-source counterparts in terms of trustworthiness, raising concerns about the potential risks of widely accessible open-source LLMs. However, a few open-source LLMs come very close to proprietary ones. Thirdly, it is important to note that some LLMs may be overly calibrated towards exhibiting trustworthiness, to the extent that they compromise their utility by mistakenly treating benign prompts as harmful and consequently not responding. Finally, we emphasize the importance of ensuring transparency not only in the models themselves but also in the technologies that underpin trustworthiness. Knowing the specific trustworthy technologies that have been employed is crucial for analyzing their effectiveness.

Russia and Ukraine traded deadly aerial attacks on civilian centres in the past week of the war, but Ukraine also scored hits on military and economic infrastructure deep in the Russian heartland, extending its reach to St Petersburg for the first time. Ukrainian military intelligence said it had struck an unspecified military target in St Petersburg on Thursday, using drones launched from Ukrainian soil. Ukrainian strategic industries minister Oleksandr Kamyshin confirmed the attack, telling the World Economic Forum in Davos that the attack was carried out by a Ukrainian-built drone that had travelled 1,250km (780 miles) from Ukrainian soil. Russia's defence ministry said three drones had been launched and it had downed all three over the Gulf of Finland that day, one near an oil terminal. On Sunday, Ukraine attacked again in several locations, and this time, the evidence of its success was clear.

Document-level Relation Triplet Extraction (DocRTE) is a fundamental task in information systems that aims to simultaneously extract entities with semantic relations from a document. Existing methods heavily rely on a substantial amount of fully labeled data. However, collecting and annotating data for newly emerging relations is time-consuming and labor-intensive. Recent advanced Large Language Models (LLMs), such as ChatGPT and LLaMA, exhibit impressive long-text generation capabilities, inspiring us to explore an alternative approach for obtaining auto-labeled documents with new relations. In this paper, we propose a Zero-shot Document-level Relation Triplet Extraction (ZeroDocRTE) framework, which generates labeled data by retrieval and denoising knowledge from LLMs, called GenRDK. Specifically, we propose a chain-of-retrieval prompt to guide ChatGPT to generate labeled long-text data step by step. To improve the quality of synthetic data, we propose a denoising strategy based on the consistency of cross-document knowledge. Leveraging our denoised synthetic data, we proceed to fine-tune the LLaMA2-13B-Chat for extracting document-level relation triplets. We perform experiments for both zero-shot document-level relation and triplet extraction on two public datasets. The experimental results illustrate that our GenRDK framework outperforms strong baselines.