We present Charles University submissions to the WMT22 General Translation Shared Task on Czech-Ukrainian and Ukrainian-Czech machine translation. We present two constrained submissions based on block back-translation and tagged back-translation and experiment with rule-based romanization of Ukrainian. Our results show that the romanization only has a minor effect on the translation quality. Further, we describe Charles Translator, a system that was developed in March 2022 as a response to the migration from Ukraine to the Czech Republic. Compared to our constrained systems, it did not use the romanization and used some proprietary data sources.

We present the CUNI-Bergamot submission for the WMT22 General translation task. We compete in English$\rightarrow$Czech direction. Our submission further explores block backtranslation techniques. Compared to the previous work, we measure performance in terms of COMET score and named entities translation accuracy. We evaluate performance of MBR decoding compared to traditional mixed backtranslation training and we show a possible synergy when using both of the techniques simultaneously. The results show that both approaches are effective means of improving translation quality and they yield even better results when combined.

本文概述了与CRAC 2022研讨会相关的多语言核心分辨率的共享任务。共同的任务参与者应该开发能够识别提及并根据身份核心重点聚集的训练系统。Corefud 1.0的公共版本包含10种语言的13个数据集，被用作培训和评估数据的来源。先前面向核心共享任务中使用的串联分数用作主要评估度量。5个参与团队提交了8个核心预测系统；此外，组织者在共享任务开始时提供了一个基于竞争变压器的基线系统。获胜者系统的表现优于基线12个百分点（就所有语言的所有数据集而言，在所有数据集中平均得分）。

已知深神经模型对输入噪声的敏感性是一个具有挑战性的问题。在NLP中，模型性能通常与自然发生的噪声恶化，例如拼写错误。要缓解此问题，模型可能会利用人为中断数据。然而，到目前为止已经任意确定产生的噪声的量和类型。因此，我们建议统计从语法纠错的语料库统计上的错误。我们对多种语言的若干先进的NLP系统进行了彻底的评估，其中任务包括句法分析，名为实体识别，神经机翻译，胶水基准和阅读理解的子集。我们还比较两种解决性能下降的方法：a）培训我们框架生成的中断数据的NLP模型;b）减少外部系统进行自然语言校正的输入噪声。代码在https://github.com/ufal/kazitext上发布。

The performance of the Deep Learning (DL) models depends on the quality of labels. In some areas, the involvement of human annotators may lead to noise in the data. When these corrupted labels are blindly regarded as the ground truth (GT), DL models suffer from performance deficiency. This paper presents a method that aims to learn a confident model in the presence of noisy labels. This is done in conjunction with estimating the uncertainty of multiple annotators. We robustly estimate the predictions given only the noisy labels by adding entropy or information-based regularizer to the classifier network. We conduct our experiments on a noisy version of MNIST, CIFAR-10, and FMNIST datasets. Our empirical results demonstrate the robustness of our method as it outperforms or performs comparably to other state-of-the-art (SOTA) methods. In addition, we evaluated the proposed method on the curated dataset, where the noise type and level of various annotators depend on the input image style. We show that our approach performs well and is adept at learning annotators' confusion. Moreover, we demonstrate how our model is more confident in predicting GT than other baselines. Finally, we assess our approach for segmentation problem and showcase its effectiveness with experiments.

Landing an unmanned aerial vehicle unmanned aerial vehicle (UAV) on top of an unmanned surface vehicle (USV) in harsh open waters is a challenging problem, owing to forces that can damage the UAV due to a severe roll and/or pitch angle of the USV during touchdown. To tackle this, we propose a novel model predictive control (MPC) approach enabling a UAV to land autonomously on a USV in these harsh conditions. The MPC employs a novel objective function and an online decomposition of the oscillatory motion of the vessel to predict, attempt, and accomplish the landing during near-zero tilt of the landing platform. The nonlinear prediction of the motion of the vessel is performed using visual data from an onboard camera. Therefore, the system does not require any communication with the USV or a control station. The proposed method was analyzed in numerous robotics simulations in harsh and extreme conditions and further validated in various real-world scenarios.

We develop theory and methods that use the graph Laplacian to analyze the geometry of the underlying manifold of point clouds. Our theory provides theoretical guarantees and explicit bounds on the functional form of the graph Laplacian, in the case when it acts on functions defined close to singularities of the underlying manifold. We also propose methods that can be used to estimate these geometric properties of the point cloud, which are based on the theoretical guarantees.

Nearly all jurisdictions in the United States require a professional license exam, commonly referred to as "the Bar Exam," as a precondition for law practice. To even sit for the exam, most jurisdictions require that an applicant completes at least seven years of post-secondary education, including three years at an accredited law school. In addition, most test-takers also undergo weeks to months of further, exam-specific preparation. Despite this significant investment of time and capital, approximately one in five test-takers still score under the rate required to pass the exam on their first try. In the face of a complex task that requires such depth of knowledge, what, then, should we expect of the state of the art in "AI?" In this research, we document our experimental evaluation of the performance of OpenAI's `text-davinci-003` model, often-referred to as GPT-3.5, on the multistate multiple choice (MBE) section of the exam. While we find no benefit in fine-tuning over GPT-3.5's zero-shot performance at the scale of our training data, we do find that hyperparameter optimization and prompt engineering positively impacted GPT-3.5's zero-shot performance. For best prompt and parameters, GPT-3.5 achieves a headline correct rate of 50.3% on a complete NCBE MBE practice exam, significantly in excess of the 25% baseline guessing rate, and performs at a passing rate for both Evidence and Torts. GPT-3.5's ranking of responses is also highly-correlated with correctness; its top two and top three choices are correct 71% and 88% of the time, respectively, indicating very strong non-entailment performance. While our ability to interpret these results is limited by nascent scientific understanding of LLMs and the proprietary nature of GPT, we believe that these results strongly suggest that an LLM will pass the MBE component of the Bar Exam in the near future.

The future of population-based breast cancer screening is likely personalized strategies based on clinically relevant risk models. Mammography-based risk models should remain robust to domain shifts caused by different populations and mammographic devices. Modern risk models do not ensure adaptation across vendor-domains and are often conflated to unintentionally rely on both precursors of cancer and systemic/global mammographic information associated with short- and long-term risk, respectively, which might limit performance. We developed a robust, cross-vendor model for long-term risk assessment. An augmentation-based domain adaption technique, based on flavorization of mammographic views, ensured generalization to an unseen vendor-domain. We trained on samples without diagnosed/potential malignant findings to learn systemic/global breast tissue features, called mammographic texture, indicative of future breast cancer. However, training so may cause erratic convergence. By excluding noise-inducing samples and designing a case-control dataset, a robust ensemble texture model was trained. This model was validated in two independent datasets. In 66,607 Danish women with flavorized Siemens views, the AUC was 0.71 and 0.65 for prediction of interval cancers within two years (ICs) and from two years after screening (LTCs), respectively. In a combination with established risk factors, the model's AUC increased to 0.68 for LTCs. In 25,706 Dutch women with Hologic-processed views, the AUCs were not different from the AUCs in Danish women with flavorized views. The results suggested that the model robustly estimated long-term risk while adapting to an unseen processed vendor-domain. The model identified 8.1% of Danish women accounting for 20.9% of ICs and 14.2% of LTCs.

Large language models (LLMs) have demonstrated impressive capabilities in natural language understanding and generation, but the quality bar for medical and clinical applications is high. Today, attempts to assess models' clinical knowledge typically rely on automated evaluations on limited benchmarks. There is no standard to evaluate model predictions and reasoning across a breadth of tasks. To address this, we present MultiMedQA, a benchmark combining six existing open question answering datasets spanning professional medical exams, research, and consumer queries; and HealthSearchQA, a new free-response dataset of medical questions searched online. We propose a framework for human evaluation of model answers along multiple axes including factuality, precision, possible harm, and bias. In addition, we evaluate PaLM (a 540-billion parameter LLM) and its instruction-tuned variant, Flan-PaLM, on MultiMedQA. Using a combination of prompting strategies, Flan-PaLM achieves state-of-the-art accuracy on every MultiMedQA multiple-choice dataset (MedQA, MedMCQA, PubMedQA, MMLU clinical topics), including 67.6% accuracy on MedQA (US Medical License Exam questions), surpassing prior state-of-the-art by over 17%. However, human evaluation reveals key gaps in Flan-PaLM responses. To resolve this we introduce instruction prompt tuning, a parameter-efficient approach for aligning LLMs to new domains using a few exemplars. The resulting model, Med-PaLM, performs encouragingly, but remains inferior to clinicians. We show that comprehension, recall of knowledge, and medical reasoning improve with model scale and instruction prompt tuning, suggesting the potential utility of LLMs in medicine. Our human evaluations reveal important limitations of today's models, reinforcing the importance of both evaluation frameworks and method development in creating safe, helpful LLM models for clinical applications.