Category: Publishing & media (Page 1 of 53)

Elsevier, provider of advanced information and decision support, launched ScienceDirect AI, a generative AI tool for researchers. It is designed to transform the way researchers work by enabling them to instantly extract, summarize and compare trusted insights from millions of full-text articles and book chapters on ScienceDirect, a platform for trusted, peer-reviewed research.

Users can search and get answers from within the full-text of 14 million articles and book chapters, using their own words to describe what they need and why. ScienceDirect AI searches across the millions of documents in its index to provide a Summary Response with references, Source Snippets for each reference, and short Related Insights summaries while linking back to the original document. A conversational feature answers questions about the content of a specific full-text article or book chapter and allows researchers to ask further questions.

ScienceDirect AI’s Compare Experiments tool takes a set of articles and creates a table breaking down each experiment within them, drawing out aspects of each including goals, methods and results.

ScienceDirect AI’s use of third-party LLMs is private, no information is stored or used to train public models, and all data is stored in a protected and private environment exclusive to Elsevier.

https://elsevier.shorthandstories.com/sciencedirect-ai

Foxit, a provider of PDF and eSignature products and services, helping knowledge workers to increase productivity and do more with documents, today announced the addition of multi-document analysis to its AI Assistant feature within its mobile apps for iOS and Android. This new and unique capability enhances the AI Assistant’s functionality, enabling users to analyze multiple documents simultaneously directly from their mobile devices.

The introduction of multi-document analysis caters to the evolving needs of professionals who require efficient tools to manage and process large volumes of information. The new feature is designed to simplify workflows by allowing users to:

• Summarize Multiple Documents at Once: Create concise summaries for all uploaded documents, saving time by avoiding the need to review each individually.
• Ask Questions Across Files: Get answers that draw insights from multiple documents simultaneously, ideal for comparing contracts, reports, or research materials.
• Extract Key Information: Pinpoint critical data, such as deadlines or figures, or find patterns and trends from across multiple files in seconds.

The multi-document analysis feature is now available in the latest update to Foxit’s PDF Reader and PDF Editor mobile apps for iOS and Android.

https://foxit.com

Version 3.0 increases the efficiency of using the tool as certain actions now leverage the Retrieval-Augmented Generation (RAG) process to obtain context from the users’ current projects.

The new AI Positron Assistant drop-down widget offers a convenient way of accessing useful AI actions by displaying a floating contextual menu directly within the editing area. Users can customize their own AI actions to display as Quick Assist fixes in the editor. It is also now possible to choose the OpenAI model used in chat sessions and actions right from the AI Positron Assistant view.

A variety of new AI actions that are specific to working with DITA XML documents have been implemented, including a Proofread action that helps users identify potential issues in their content regarding logical consistency, grammar, spelling, readability, and comprehension.

Other newly implemented actions include the Improve Structure action that instructs the AI to enhance DITA XML documents by adding additional structure or inline elements, and the Add Structured Content action continues the content from a document with additional structured content generated based upon similar content from the current project, which gives the AI more context for formulating the new XML structure.

 https://www.oxygenxml.com/ai_positron_assistant.html

DeepL, a global Language AI company, announced its next-generation language model, powered by a highly-specialized LLM technology built specifically for translation and editing to outperform competitors. The update enhances translation quality and performance in DeepL’s Language AI platform for businesses:

• A specialized LLM: the DeepL solution leverages an LLM uniquely tuned for language, resulting in more human-like translations and writing for a variety of use cases with a reduced risk of hallucinations and misinformation.
• Proprietary data: unlike general purpose models that simply train on the public internet, the DeepL model leverages over seven years of proprietary data specifically tuned for content creation and translation.
• Human model tutoring: with a focus on quality, the DeepL model leverages thousands of hand-picked language experts specifically trained to “tutor” the model to best-in-class translation.

Translations using the next-gen model are available for DeepL Pro customers for translations in English, Japanese, German, and Simplified Chinese, with additional languages coming soon. The LLM can be activated within the web translator by selecting “next-gen model”. DeepL Pro users are protected by enterprise-grade security and compliance standards (ISO 27001 certification, GDPR/SOC 2 type 2 compliance), and no Pro translations are ever used to train its models.

https://www.deepl.com/en/press-release#2GYyHCVU8bjbu1iewCwLLx