Natural language processing
Field of linguistics and computer science
Follow Natural language processing on Notably News to receive short updates to your email — rarely!
We include updates on Google Translate, Siri, Yandex Translate, Andrew Ng, Google Assistant, DeepL Translator, CarPlay, Prompt engineering, Retrieval-augmented generation, Android Auto, Microsoft Translator, Foundation model, Neural machine translation, Word embedding, IBM Watson, WolframAlpha ... and more.
| March 14 2025 |
Google Assistant
Google announces that Assistant will stop working on Android devices using Android 10 and higher, iOS, and other devices like Google Nest, with Gemini mostly replacing it. Assistant will remain functional on low-range phones or devices running Android 9 'Pie' and lower.
|
| February 2025 | Nvidia disclosed in a regulatory filing that it had sold its stake in SoundHound AI. |
| February 2025 | SoundHound introduced brand personalities for its Chat AI automotive voice assistant, allowing car manufacturers to create unique vehicle voice identities. |
| January 2025 |
Large language model
DeepSeek released DeepSeek R1, a 671-billion-parameter open-weight model performing comparably to OpenAI o1 at a lower cost.
|
| January 2025 |
Large language model
DeepSeek released DeepSeek-R1, a 671-billion-parameter open-weight reasoning model that achieved comparable performance to OpenAI's o1 model, offering a more cost-effective alternative with open-source accessibility.
|
| January 2025 | SoundHound announced a partnership with Rekor Systems to integrate voice and conversational AI with vehicle recognition software. |
| January 2025 | Lucid launched the Lucid Assistant powered by SoundHound Chat AI. |
| January 2025 | At CES 2025 in Las Vegas, SoundHound introduced an in-vehicle voice commerce platform for placing takeout orders through car infotainment systems. |
| 2024 |
Google Translate
Added 110 new languages including Cantonese, Tok Pisin, and several regional Russian languages (Bashkir, Chechen, Ossetian, and Crimean Tatar) using the PaLM 2 AI model.
|
| 2024 |
Pronunciation assessment
Development of promising improvements in pronunciation assessment, including articulatory feature extraction and transfer learning techniques to reduce unnecessary corrections.
|
| 2024 |
Large language model
Researcher Yongge Wang demonstrated a potential method to bypass ChatGPT 4o's safety controls, showing how a malicious actor could potentially obtain information about establishing a drug trafficking operation.
|
| 2024 |
Large language model
OpenAI released the reasoning model OpenAI o1, which generates long chains of thought before providing answers.
|
| 2024 | Awarded Honorary Fellowship of the Royal Statistical Society |
| December 2024 |
Large language model
OpenAI launched the o3 model, continuing their series of reasoning models with enhanced capabilities for step-by-step problem solving.
|
| November 2024 | SoundHound integrated its voice AI into Kia vehicles in India, including support for the Hindi language. |
| October 2024 | AI Fund made its first investment in India, backing AI healthcare startup Jivi, which uses AI for diagnoses, treatment recommendations, and administrative tasks. |
| October 2024 | SoundHound won the 'Overall Connected Solution of the Year' award in the AutoTech Breakthrough Awards Program for its Chat AI solution. |
| September 2024 | Governor Gavin Newsom vetoed the California AI legislation that Ng had previously criticized, which would have imposed significant regulatory requirements on AI model developers. |
| September 2024 |
Large language model
OpenAI introduced the o1 model, a new type of reasoning model designed for complex problem-solving tasks, marking a significant advancement in large language model development with improved performance in mathematics, science, and coding.
|
| September 2024 |
Large language model
Mistral introduced its multimodal Pixtral 12B model, expanding the capabilities of large language models to process multiple input types.
|
| August 2024 | SoundHound acquired Amelia AI for $85 million to increase presence in financial services, insurance, healthcare, retail, and hospitality sectors. |
| July 2024 |
Bixby
Samsung confirmed plans to launch an upgraded version of Bixby powered by its proprietary large language model (LLM) technology, promising enhanced capabilities through generative AI.
|
| June 2024 | Ng critiqued proposed AI legislation in California in a Financial Times interview, describing a bill with proposed safety mechanisms like a 'kill switch' for advanced AI models as creating 'massive liabilities for science-fiction risks'. |
| June 2024 |
Google Translate
Google Translate reached its 54th stage, dramatically expanding language support to include 100+ additional languages, including regional variations, and adding speech programs for languages like Amharic, Cantonese, and Welsh.
|
| June 2024 | SoundHound acquired Allset, a restaurant ordering marketplace founded out of Ukraine. |
| June 2024 |
Pronunciation assessment
Audio multimodal large language models begin to be used for pronunciation assessment.
|
| June 2024 |
Pronunciation assessment
Emergence of augmented reality interfaces for mobile devices that use optical character recognition to provide pronunciation training based on text found in user environments.
|
| April 11 2024 | Appointed to Amazon's board of directors. |
| March 28 2024 |
Google Translate
Google officially closed the Translate Community program, ending its crowdsourcing efforts to improve translation quality across languages.
|
| February 2024 |
Large language model
Google presented the Gemini 1.5 large language model, which can handle a context window of up to 1 million tokens, with a context window of 10 million tokens successfully tested.
|
| February 8 2024 |
Google Assistant
The Gemini chatbot is released in the United States, marking a significant step in Google's AI assistant evolution.
|
| 2023 |
IBM Watson
Watson results in IBM losing 10% of its stock value and causing mass layoffs, highlighting the project's commercial failure.
|
| 2023 |
IBM Watson
IBM shifted focus to WatsonX, a new initiative aimed at developing narrower, industry-targeted technology within IBM's cloud computing platform strategies.
|
| 2023 | SoundHound's Dynamic Interaction solution was selected as the 'Best Use of AI' at the Automation & Self-Service Awards. |
| 2023 |
Pronunciation assessment
Researchers developed a method to assess intelligibility using dynamic time warping based distance from Wav2Vec2 representation of good speech.
|
| 2023 |
RiTa
RiTa released Version 3.0, implementing the software toolkit in both Java and JavaScript, with integrations for p5.js, Processing, Node, and Android platforms.
|
| 2023 |
Large language model
A Microsoft team argued that GPT-4 could be viewed as an early version of an artificial general intelligence system, capable of solving novel and difficult tasks across multiple domains.
|
| 2023 |
Large language model
Goldman Sachs projected that generative language AI could potentially increase global GDP by 7% within the next decade and could potentially automate around 300 million jobs worldwide.
|
| 2023 |
Large language model
Nature Biomedical Engineering published a statement highlighting the indistinguishability of human-written and AI-generated text, predicting the rapid proliferation of large language models and their potential to transform multiple industries.
|
| 2023 |
Large language model
Training a 12-billion-parameter LLM requires approximately 72,300 A100-GPU-hours, showcasing continued improvements in computational efficiency.
|
| 2023 |
Large language model
GPT-4 was released, praised for increased accuracy and multimodal capabilities. Many LLMs began training to become multimodal, processing various data types.
|
| 2023 |
Large language model
A research study demonstrated ChatGPT 3.5 turbo's unique memorization behavior, where repeated prompting to output the same word would eventually trigger verbatim excerpts from its training data after hundreds of repetitions.
|
| 2023 |
Large language model
A research study proposed a method for circumventing large language model (LLM) safety systems, highlighting vulnerabilities in existing content filtering mechanisms.
|
| 2023 | Named in Time AI 100 Most Influential People |
| 2023 | Expanded access to AI education, with an estimated 8 million individuals worldwide taking his courses via platforms like DeepLearning.AI and Coursera. |
| December 2023 | SoundHound acquired SYNQ3 Restaurant Solutions, a provider of voice AI for restaurants, for $25 million. |
| December 2023 | Provided an interview to Financial Times discussing AI regulation, highlighting concerns about potential negative impacts on open-source AI development and innovation. |
| October 2023 |
Google Assistant
A mobile version of the Gemini chatbot, initially titled Assistant with Bard and later just Bard, is unveiled during the Pixel 8 event, signaling a potential replacement for Google Assistant.
|
| June 2023 |
Google Translate
Google Translate launched its 53rd stage, adding speech programs for Lithuanian and Punjabi languages.
|
|
We are only showing the most recent entries for this topic. |
|
This contents of the box above is based on material from the Wikipedia articles Pronunciation assessment, Large language model, Bixby (software), Andrew Ng, Google Books Ngram Viewer, RiTa, SoundHound, Google Assistant, IBM Watson & Google Translate, which are released under the Creative Commons Attribution-ShareAlike 4.0 International License.