AmazonScience/document-haystack
Document Haystack Dataset This repository contains the dataset for the paper “Document Haystack: A Long Context Multimodal Image/Document Understanding Vision LLM Benchmark”. 📑 Abstract Paper The proliferation of multimodal Large Language Models has significantly advanced the ability to analyze and understand complex data inputs from different modalities. However, the processing of long documents remains under-explored, largely due to a lack of suitable… See the full description on the dataset page: https://huggingface.co/datasets/AmazonScience/document-haystack.
2090k
1A PROACTIVE2Both our business and more broadly our 3sector are constantly evolving and each year 4we are faced with new challenges. However, 5we consistently prove ourselves adept at 6weathering difficulties, delivering against 7our commitments and adapting to change.8Resolving past uncertainties9In 2023 we made significant progress in 10resolving two long-standing issues, both 11with material benefits. Firstly we were able to 12reach agreement on our outstanding pension 13valuation with the MGN pension scheme, 14avoiding costly regulatory intervention and 15providing clarity that these financial obligations 16will in the main unwind in early 2028.17After a lengthy legal process we have also 18been able to achieve clarity around our 19historical legal issues. December’s judgment 20on time limitation has materially reduced 21our expected obligations and, barring 22exceptional circumstances, brought 23an end to any future claims.24↓ £20M25Estimated reduction in 26historical legal issue costs27↓ c.£40M28Estimated reduction in 29pension obligations in 20283043%31Data-driven revenues3211%33RPM increase34Moving forward to digital-first35In 2023 we delivered a 5.7% reduction in 36operating costs (on a like-for-like basis) 37and in November announced a similar 5-6% 38reduction for 2024. These savings decisions, 39while never easy, are made to support the 40future of our business.41A guiding principle behind these changes was 42the need to more firmly orientate our newsrooms 43and wider organisation towards our digital 44audience. This meant considering online 45behaviour in all of our decisions – topics, 46timing, format – and rethinking how we tell 47every story in today’s digital landscape.48Initiatives include the automation of our 49content management system (CMS) so 50journalists can save time uploading stories, 51sharing more content across brands, and 52organising teams for maximum impact. 53For example we have brought together 54our video and audio talent into one Studio 55team, to produce content for both our 56editorial brands and commercial partners 57and to better support the branded content 58revenue opportunity. 59APPROACH601061Reach plc Annual Report 2023Strategic Report Governance Financial Statements Other Information