We introduce DeepPSL a variant of probabilistic soft logic (PSL) to produce an end-to-end trainable system that integrates reasoning and perception. PSL represents first-order logic in terms of a convex graphical model – hinge-loss Markov random fields (HL-MRFs). PSL stands out among probabilistic logic frameworks due to its tractability having been applied to systems of more than 1 billion ground rules. The key to our approach is to represent predicates in first-order logic using deep neural networks and then to approximately back-propagate through the HL-MRF and thus train every aspect of the first-order system being represented. We believe that this approach represents an interesting direction for the integration of deep learning and reasoning techniques with applications to knowledge base learning, multi-task learning, and explainability. Evaluation on three different tasks demonstrates that DeepPSL significantly outperforms state-of-the-art neuro-symbolic methods on scalability while achieving comparable or better accuracy.

Removing noise from scanned pages is a vital step before their submission to optical character recognition (OCR) system. Most available image denoising methods are supervised where the pairs of noisy/clean pages are required. However, this assumption is rarely met in real settings. Besides, there is no single model that can remove various noise types from documents. Here, we propose a unified end-to-end unsupervised deep learning model, for the first time, that can effectively remove multiple types of noise, including salt & pepper noise, blurred and/or faded text, as well as watermarks from documents at various levels of intensity. We demonstrate that the proposed model significantly improves the quality of scanned images and the OCR of the pages on several test datasets.

In the regular course of business, companies spend a lot of effort reading and interpreting documents, a highly manual process that involves tedious tasks, such as identifying dates and names or locating the presence or absence of certain clauses in a contract. Dealing with natural language is complex and further complicated by the fact that these documents come in various formats (scanned image, digital formats) and have different degrees of internal structure (spreadsheets, invoices, text documents). We present DICR, an end-to-end, modular, and trainable system that automates the mundane aspects of document review and allows humans to perform the validation. The system is able to speed up this work while increasing quality of information extracted, consistency, throughput, and decreasing time to decision. Extracted data can be fed into other downstream applications (from dashboards to Q&A and to report generation).