Legal Judgment Prediction is one of the most acclaimed fields for the combined area of NLP, AI, and Law. By legal prediction we mean an intelligent systems capable to predict specific judicial characteristics, such as judicial outcome, a judicial class, predict an specific case. In this research, we have used AI classifiers to predict judicial outcomes in the Brazilian legal system. For this purpose, we developed a text crawler to extract data from the official Brazilian electronic legal systems. These texts formed a dataset of second-degree murder and active corruption cases. We applied different classifiers, such as Support Vector Machines and Neural Networks, to predict judicial outcomes by analyzing textual features from the dataset. Our research showed that Regression Trees, Gated Recurring Units and Hierarchical Attention Networks presented higher metrics for different subsets. As a final goal, we explored the weights of one of the algorithms, the Hierarchical Attention Networks, to find a sample of the most important words used to absolve or convict defendants.

Remote sensing of the Earth's surface water is critical in a wide range of environmental studies, from evaluating the societal impacts of seasonal droughts and floods to the large-scale implications of climate change. Consequently, a large literature exists on the classification of water from satellite imagery. Yet, previous methods have been limited by 1) the spatial resolution of public satellite imagery, 2) classification schemes that operate at the pixel level, and 3) the need for multiple spectral bands. We advance the state-of-the-art by 1) using commercial imagery with panchromatic and multispectral resolutions of 30 cm and 1.2 m, respectively, 2) developing multiple fully convolutional neural networks (FCN) that can learn the morphological features of water bodies in addition to their spectral properties, and 3) FCN that can classify water even from panchromatic imagery. This study focuses on rivers in the Arctic, using images from the Quickbird, WorldView, and GeoEye satellites. Because no training data are available at such high resolutions, we construct those manually. First, we use the RGB, and NIR bands of the 8-band multispectral sensors. Those trained models all achieve excellent precision and recall over 90% on validation data, aided by on-the-fly preprocessing of the training data specific to satellite imagery. In a novel approach, we then use results from the multispectral model to generate training data for FCN that only require panchromatic imagery, of which considerably more is available. Despite the smaller feature space, these models still achieve a precision and recall of over 85%. We provide our open-source codes and trained model parameters to the remote sensing community, which paves the way to a wide range of environmental hydrology applications at vastly superior accuracies and 2 orders of magnitude higher spatial resolution than previously possible.

Telcos are hoping that 5G will transform enterprises the way 4G helped consumers. For instance, watching Fifa World Cup 2022 would have been a delight for many with 5G networks. The first rollouts took place in 2019 in South Korea and the US, but since then 5G has been rolled out in over 70 countries. In India, 5G services were launched by Bharti Airtel and Reliance Jio in October. According to a GSMA Intelligence report, published in October, 5G can contribute $455 billion to the Indian economy between 2023 and 2040.

More than once this year, AI experts have repeated a familiar refrain: "Please slow down." AI news in 2022 has been rapid-fire and relentless; the moment you knew where things currently stood in AI, a new paper or discovery would make that understanding obsolete. In 2022, we arguably hit the knee of the curve when it came to generative AI that can produce creative works made up of text, images, audio, and video. This year, deep-learning AI emerged from a decade of research and began making its way into commercial applications, allowing millions of people to try out the tech for the first time. AI creations inspired wonder, created controversies, prompted existential crises, and turned heads.

It’s an age-old problem in procurement: Corporate buyers lack the time to negotiate fully with all suppliers. Historically this has left untapped value on the table for both buyers and suppliers. To address this challenge, Walmart deployed AI-powered negotiations software with a text-based interface (i.e., a chatbot) to connect with suppliers. So far, the chatbot is negotiating and closing agreements with 68% of suppliers approached, with each side gaining something it values. This article offers four lessons to deliver results from automated procurement negotiations: move quickly to a production pilot, start with indirect spend categories with pre-approved suppliers, decide on acceptable negotiation trade-offs, and scale by extending geographies, categories, and use cases.

Mapping crops using remote sensing technology is important for food security and land management. Machine learning-based methods has become a popular approach for crop mapping in recent years. However, the key to machine learning, acquiring ample and accurate samples, is usually time-consuming and laborious. To solve this problem, a crop mapping method in the small sample/no sample case that integrating domain knowledge and using a cascaded classification framework that combine a weak classifier learned from samples with strong features and a strong classifier trained by samples with weak feature was proposed. First, based on the domain knowledge of various crops, a low-capacity classifier such as decision tree was applied to acquire those pixels with distinctive features and complete observation sequences as "strong feature" samples. Then, to improve the representativeness of these samples, sample augmentation strategy that artificially remove the observations of "strong feature" samples according to the average valid observation proportion in target area was applied. Finally, based on the original samples and augmented samples, a large-capacity classifier such as random forest was trained for crop mapping. The method achieved an overall accuracy of 82% in the MAP crop recognition competition held by Syngenta Group, China in 2021 (third prize, ranked fourth). This method integrates domain knowledge to overcome the difficulties of sample acquisition, providing a convenient, fast and accurate solution for crop mapping.

One of the most impressive results of recent NLP history is the ability of pre-trained language models to solve new tasks in a zero-shot setting. To achieve this, NLP tasks are framed as natural language prompts, generating a response indicating the predicted output. Nonetheless, the performance in such settings often lags far behind its supervised counterpart, suggesting a large space for potential improvement. In this paper, we explore methods to utilize unlabeled data to improve zero-shot performance. Specifically, we take advantage of the fact that multiple prompts can be used to specify a single task, and propose to regularize prompt consistency, encouraging consistent predictions over this diverse set of prompts. Our method makes it possible to fine-tune the model either with extra unlabeled training data, or directly on test input at inference time in an unsupervised manner. In experiments, our approach outperforms the state-of-the-art zero-shot learner, T0 (Sanh et al., 2022), on 9 out of 11 datasets across 4 NLP tasks by up to 10.6 absolute points in terms of accuracy. The gains are often attained with a small number of unlabeled examples.

Self-supervised learning methods are gaining increasing traction in computer vision due to their recent success in reducing the gap with supervised learning. In natural language processing (NLP) self-supervised learning and transformers are already the methods of choice. The recent literature suggests that the transformers are becoming increasingly popular also in computer vision. So far, the vision transformers have been shown to work well when pretrained either using a large scale supervised data or with some kind of co-supervision, e.g. in terms of teacher network. These supervised pretrained vision transformers achieve very good results in downstream tasks with minimal changes. In this work we investigate the merits of self-supervised learning for pretraining image/vision transformers and then using them for downstream classification tasks. We propose Self-supervised vIsion Transformers (SiT) and discuss several self-supervised training mechanisms to obtain a pretext model. The architectural flexibility of SiT allows us to use it as an autoencoder and work with multiple self-supervised tasks seamlessly. We show that a pretrained SiT can be finetuned for a downstream classification task on small scale datasets, consisting of a few thousand images rather than several millions. The proposed approach is evaluated on standard datasets using common protocols. The results demonstrate the strength of the transformers and their suitability for self-supervised learning. We outperformed existing self-supervised learning methods by large margin. We also observed that SiT is good for few shot learning and also showed that it is learning useful representation by simply training a linear classifier on top of the learned features from SiT. Pretraining, finetuning, and evaluation codes will be available under: https://github.com/Sara-Ahmed/SiT.

OpenClassrooms mentors are freelance senior professionals. They help students succeed in their training programs, through weekly video calls of maximum one hour each. To be a role model and share their industry know-how with students on a weekly basis, coaching them through their training programs. All of the educational content is created and made available through the OpenClassrooms platform. In the Data Analyst path, our students learn to analyze data and model phenomena with realistic business cases.

As text generated by large language models proliferates, it becomes vital to understand how humans engage with such text, and whether or not they are able to detect when the text they are reading did not originate with a human writer. Prior work on human detection of generated text focuses on the case where an entire passage is either human-written or machine-generated. In this paper, we study a more realistic setting where text begins as human-written and transitions to being generated by state-of-the-art neural language models. We show that, while annotators often struggle at this task, there is substantial variance in annotator skill and that given proper incentives, annotators can improve at this task over time. Furthermore, we conduct a detailed comparison study and analyze how a variety of variables (model size, decoding strategy, fine-tuning, prompt genre, etc.) affect human detection performance. Finally, we collect error annotations from our participants and use them to show that certain textual genres influence models to make different types of errors and that certain sentence-level features correlate highly with annotator selection. We release the RoFT dataset: a collection of over 21,000 human annotations paired with error classifications to encourage future work in human detection and evaluation of generated text.