Benchmarking for Translation
Profound Benchmarking and Reporting of Translation Services made simple
Los Gatos, 20th January 2022
Today aiXplain launches benchmarking and reporting for translation services. This release provides member access to comprehensive machine translation (MT) model performance evaluation and reporting tools, utilizing objective industry standard scoring metrics. With aiXplain, profound model analysis is now readily available to all its members; AI consumers, suppliers, program managers and machine learning specialists alike.
Machine translation is a thriving and rapidly evolving domain, populated by numerous service providers competing to become the standard in translation performance. Due to the fast-paced nature of AI development, combined with an ever increasing demand for translation of widening content types in various languages, dialects and domains, it has become an exceedingly difficult task to keep tabs on MT model advancement.
Solving this in-house poses a daunting challenge to MT supplier’s benchmarking development efforts. Implementing the disparate pieces required to achieve meaningful machine translation metrics demands a burdensome resource allocation to achieve meaningful results, even with expert subject knowledge. Ultimately, building such a benchmarking and reporting infrastructure is an added burden and cost to most enterprises, and rarely serves as a cost effective expenditure.
With aiXplain’s benchmarking for translation and reporting tools, profound industry-specific ML performance analysis and optimization insights are now available to expert and novice users alike. aiXplain’s benchmarking service also allows members to easily select and compare specific MT models, and datasets, using state-or-the-art benchmark scoring. During the model evaluation phase, members can now easily discover and compare models as defined by their individual performance and quality requirements, without the need for an institutional knowledge of traditional benchmark program management.
With report scheduling, MT service providers can also gain valuable regression insights into model development and performance metrics on a periodical basis. According to Kamer Yuksel, Principal AI Architect at aiXplain, “we’re bringing added value to MT service providers. Members with unique language-pair and domain coverage requirements can now realize improved lead generation and offer the best customer experience with the objective, and unbiased benchmark results obtained through aiXplain.”
With aiXplain’s highly detailed benchmarking reports members can stay informed on the status of their distinct language services with respect to the overall market segment. These reports consider: translation accuracy, speed and availability on generic and niche translation, incorporating several language pairs on publicly available or proprietary MT datasets. Mudar Yaghi, CEO of AppTek highlights that, “The report delivers actionable insights used to prioritize scientific and engineering efforts while allowing our business development team to better focus efforts to match our market strengths. We rely on aiXplain’s periodic benchmarking for translation to free up resources from time-consuming benchmarking tasks and help us make smarter business decisions across the board.”
About aiXplain, Inc.
aiXplain is the place where nothing stands between you and the power of AI. We provide unique and insightful benchmarking resources, developed by highly experienced leaders in AI science, to serve suppliers, researchers, and customers in the human language technology fields. Our mission is to truly democratize AI by making it accessible to members at every stage of business development.