How NeuralSpace helps developers build language AI models in Oriya, German, and many other local languages

A few years ago, Ayushmann Dash and Felix Laumann realized that the existing natural language processing (NLP) models worked poorly in locally spoken languages ​​and the model translations weren’t really accurate.

Although NLP models existed in German (Felix’s mother tongue) and Oriya (Ayushman’s mother tongue) in 2018-2019, they did not encounter models based on the latest learning advances in depth.

So, the duo decided to create NeuralSpace, a software-as-a-service (SaaS) platform that offers developers a web interface and API suite for text and voice NLP tasks (language AI models) in more than 100 languages.

Local languages ​​and improving accuracy in real-world use cases are priority areas for the platform, which received an initial investment of $2.8 million from Techstars, a global investment firm based in Colorado.

“The Googles, Microsofts, IBMs, and AWSs of this world also offer models in these languages, but they are often inaccurate in many real-world use cases. NeuralSpace develops models with a local language-first approach. We make sure that our models perform well in locally spoken languages ​​such as Gujarati, Telugu and Gulf Arabic even before training them in English, French and Spanish,” Felix explains.

What does the platform offer?

The NeuralSpace platform consists of various language processing features in over 100 languages. It offers a suite of APIs for linguistic AI models, with features such as entity recognition, language understanding, text-to-speech, text-to-speech, machine translation, and transliteration.

Features are pre-trained with a basic understanding of each language and can be used across many industries. Developers can either consume the content directly or customize it to get their own unique AI model that precisely solves their business problems.

For example, in conversational AI solutions, such as chatbots and voicebots, the “language understanding” offering helps recognize customer intent and extract relevant information from what is said or written. to perform a relevant action. Natural language understanding can also predict user sentiment, which can help analyze user behavior. This makes it possible to create chatbots and voicebots that respond to the user in a more empathetic tone.

“For example, if the user says, ‘I want to reserve a table at 8 p.m. tonight’, our understanding of the language would recognize that the customer wants to make a reservation (intent) along with the time (8 p.m.) and date (today)”, explains Félix.

NeuralSpace can automatically detect the language a person is writing in, translate between languages, transcribe what was said into text, identify speakers in a meeting, and create a synthetic voice that reads text.

The AI ​​model that powers the platform is state-of-the-art and was developed in-house with some of the top researchers in the field, Felix says.

The technology is specially developed for languages ​​spoken locally in India, the Middle East and Southeast Asia.

The team

NeuralSpace is headquartered in London, UK, and has offices in India and Tunisia. The startup team is made up of 22 engineers, marketers and operators, all driven by a passion for breaking down language barriers on the internet.

The founders met while studying in Europe. Ayushman was studying computer science at the Technical University of Kaiserslautern in Germany, while Felix was at the Technical University of Denmark in Copenhagen.

“We greatly benefit from having 19 out of 22 employees in India, whose mother tongues are Punjabi, Bengali, Oriya, Marathi and Hindi. We also have an Arabic linguist working with us and we look forward to hiring an even more international team,” says Felix.

growth path

The NLP startup recently joined Colarado-based Techstars, a global investment platform that provides access to capital and one-on-one mentoring, and received an initial investment of $2.8 million.

The company’s product is online today and is used by over 300 companies. NeuralSpace’s top customers include Hello Ebbot, a conversational AI company for Scandinavian languages, and Eden AI, a marketplace for easily integrating and comparing various machine learning vendors.

The startup offers a pay-as-you-go revenue model.

Market and competition

The global text-to-speech market was valued at $2 billion in 2020 and is expected to reach $5 billion by 2026, according to a report by MarketsandMarkets.

NeuralSpace competes with Bengaluru-based Slang Lab, which develops augmented voice experiences for users. These experiences help users command their devices to perform actions without even having to touch the screen. NeuralSpace also competes with Deepdub, Neospaience, Papercup and Lovo.

The startup believes that its advantage lies in the number of languages ​​offered on its platform (more than 100 against 30 among its competitors Google AI, Microsoft Azure Cognitive Services and IBM Watson), the performance and accuracy of the models, and the fact that it offers an all-in-one platform with many language AI offerings.

“We have built a horizontally scalable platform that can automatically adapt to our customers’ different usage patterns. During the Black Friday flash sales, our e-commerce templates for language processing are heavily used,” says Felix.

Build capacity

Currently, the NeuralSpace team is working on expanding all of its core services, building customization capabilities and adding more languages, especially for Automatic Speech Recognition (ASR) models.

“Many people speak a mixture of two languages ​​in their daily lives. For example, “Hinglish” (Hindi-English) or a mixture of Arabic and English. We invest heavily in building models that understand two languages ​​together,” says Feliz.

In the future, the startup wants to focus on creating end-to-end solutions based on the individual services currently offered.

“For example, a voicebot is a combination of ASR, language understanding and synthetic voice generation (also known as ‘text-to-speech’). There are many more such examples and we want our customers to can create them all easily on NeuralSpace, in more than 100 languages, adapted to their specific vocabulary,” says Felix.

Source link

Comments are closed.