VALHALLA is a machine learning model developed by researchers at the Massachusetts Institute of Technology (MIT), the University of California at San Diego (UCSD), and IBM that may shake up the fields of translation, localization services, and even interpreting services.
How VALHALLA Developed in the Machine Translation Field
Imagine how you first learned languages. Did you read the words directly? Did you chat with others fluently? Absolutely not. It requires a knowledge base. In infancy, we begin to learn words by pointing and interacting with environments. Through continuous practice and accumulation, we finally gain the ability to create complete sentences.
Similarly, in the process of learning a new language, we use our senses, including hearing and vision, to learn through multimedia. We also associate new words with familiar words to improve language acquisition and retention. Then, with enough practice, we can accurately translate new, unseen sentences into context without the accompanying media. This is the basis of VALHALLA, the new machine learning model.
How VALHALLA Differs From Other Machine Translation Models
In traditional machine translation, machine learning models have worked mainly using text-based information. In recent years, however, some researchers have grown interested in multi-model machine translation, which takes more than one form of input, such as text and a corresponding image. In VALHALLA, a trained neural network uses the text input, creates an image of what it looks like, and then uses both to predict the most accurate sentence in the target language.
The VALHALLA team found that it could boost machine translation accuracy over text-only translation. Furthermore, it gave an additional boost for cases with long sentences, under-resourced languages, and instances where the machine translator is unable to access a portion of the source sentence.
The Future of Translation
The team plans to explore other means of improving language translation in the future. “Here, we only focus on images, but there are other types of a multimodal information — for example, speech, video or touch, or other sensory modalities,” says Rameswar Panda, the VALHALLA team member. “We believe such multimodal grounding can lead to even more efficient machine translation models, potentially benefiting translation across many low-resource languages spoken in the world.”
Machine translation is a common and accessible tool for document translation needs. VALHALLA is an advance in the field of machine translation. However, machine translation cannot wholly replace human translation. If your business wants to precisely translate a text’s tone, slang, culture, and idioms, you can hire professional translation services at a translation agency.
About Language Connections:
Language Connections is one of the top language service companies in the US. Over the last 30 years, we’ve focused on providing the best business translation services, interpreting services, as well as interpreter training and customized language training programs. In addition to top-tier corporate language training, we offer certified corporate interpreters and professional business translation services in 200+ languages. Our network includes linguists with backgrounds in all major industries. They’re ready to meet your needs, whether they’re for technical translation services, legal translation, government translation services, international development translation services, education translation services, life sciences translation, or something else. Reach out to us today for a free quote on our cost-efficient and timely translation services, interpreters, or other linguistic services.
Language Connections Inc.
2001 Beacon Street, Suite 105,
Boston, MA 02135