DeepPavlov

04 Oct 2022 09:29

Hello, everyone!

We are happy to share with you a new article about the #DREAM Dialogue Assistant 🔥

In this article, we highlighted the role of of Knowledge Base-specific Annotators for the utterances’ analysis within the DREAM dialogue assistant.

Follow the link to read the article, and don't hesitate to contact us if you have any questions or comments!

DeepPavlov

23 Sep 2022 08:51

Hello, everyone!

Maria Molchanova, Research Product Manager of DeepPavlov.ai, will be a speaker at the Open Data conference. Open Data is a platform where representatives of science, business and society meet every year. This event will be held this Friday in Saint-Petersburg.

Maria will be a speaker at technological track "Big Data&AI", and she will make a presentation on the topic: “DeepPavlov Dream: on the way to common datasets and conversational AI”.

Please join us and learn more about simple and complex chatbots’ development with DeepPavlov.ai ☺️

You can find out more details below:
https://opendata2021.ru/

#events #conferences #convai

DeepPavlov

09 Sep 2022 17:05

Hi, friends!
This is our regular #ConvAInews!

◾️Amazon Web Services team has made strides this summer in the CCaaS space. They have added new workforce management, call transcription, and conversational AI features to its contact center as a service platform.

◾️ Meta has introduced Velox, an open source unified execution engine aimed at accelerating data management systems and streamlining their development. Velox provides value by unifying the execution engines across different data computation engines. The three most popular integrations are Presto, Spark, and TorchArrow/PyTorch.

◾️ Yellow.ai has released DynamicNLP, a sophisticated natural language processing (NLP) solution that will power its conversational AI offerings. You can find more about it here.

◾️ Microsoft is teaching computers to understand cause and effect. Causal machine learning with Microsoft's next-best-question model could replace AB testing to help you make better business decisions. You can find more about it here.

Stay tuned!

DeepPavlov

04 Sep 2022 11:51

Hello friends,

Unfortunately, several days ago we lost access to our repositories on GitHub.
Luckily, our DeepPavlov Library and other projects we hardly worked on during the last 4 years survived.

You can find them here:

GitHub repo: https://github.com/deeppavlovteam/DeepPavlov
Gitee: https://gitee.com/organizations/mirrors_deepmipt/

We need your support as never before and kindly ask you to star us on GitHub!

Please stay tuned for updates!

DeepPavlov

12 Aug 2022 11:59

Hi, friends!
This is our regular #ConvAInews!
 
◾️Daniel Kornev, CPO of DeepPavlov.ai, will be a speaker at the PyData conference in Yerevan today at 3:30 pm (GMT+4). He will tell how to build your own Multiskill AI Assistant with DeepPavlov. Here is the link with the details. Please join us!

◾️Gizmodo made a sum-up of the last 20 years of DARPA's work. They dug into DARPA's history to suss out the agency's most significant, sometimes scary research from the last two decades.

◾️Meta has released BlenderBot 3, an advanced bot responder dataset, which can engage with humans more naturally while utilizing more prompts to guide users along a specific path of inquiry. 

◾️Google introduced a web version of its popular Read Along educational app for teaching children to read. The website includes Read Along the virtual assistant Diya, a voice-enabled AI guiding and correcting kids as they practice reading.

Stay tuned!

DeepPavlov

05 Aug 2022 13:24

Hello everyone!

Have you ever dreamed of becoming
a part of a truly massive project and changing the world?
Now is the right time to do so😊

At DeepPavlov.ai we are looking for talents - those who are ready to join our team and develop a conversational AI technology stack that helps users around the world build simple chatbots and complex AI assistants.

Here are several open positions:
• • Software Engineer (Python), DeepPavlov Dream,
• • Software Engineer (Python), DeepPavlov Agent,
• • Software Engineer (Python), Dev Tools/DF Designer,
• • Software Engineer (Frontend), Dev Tools/DF Designer,
• • Software Engineer Intern (Python, Java), Applied Research.

We are really looking forward to hearing from you!

#jobs

DeepPavlov

28 Jul 2022 17:36

❗Напоминаем, что Community Call, посвященный способу управления ходом диалога в чат-боте на основе дискурсивного анализа текста, с Борисом Галицким начнется чуть меньше, чем через полчаса, в 19:00. Подключиться можно по ссылке

DeepPavlov

27 Jul 2022 18:10

А мы начали семинар об управлении диалогом -)
Присоединяйтесь к нам по ссылке.

Также трансляцию можно посмотреть на YouTube.

DeepPavlov

25 Jul 2022 17:53

🔥Mikhail Burtsev, founder and leader of open-source conversational AI framework DeepPavlov, scientific director of the Artificial Intelligence Research Institute (AIRI), participated in a very interesting interview with Tim Ventura.

Tim Ventura is a futurist & digital marketing executive who learns from experts from all over the world everything about space, technology, science & policy.

🤖Mikhail shared with Tim his views on the future of AI. Among the topics covered:

▪️Trends in Artificial Intelligence and applications in medicine and financial markets
▪️Personal digital assistants and Google LaMDA,
▪️Passing the Turing Test and Ray Kurzweil's Technological Singularity,
▪️How open-source community software is helping drive this field forward for scientists and innovators across the globe.

🎧Follow the link to watch a video about all these aspects.

DeepPavlov

18 Jul 2022 12:16

Друзья, всем привет 👋

🔥 28 июля в 19:00 приглашенный гость Борис Галицкий, ассоциированный сотрудник лаборатории Интеллектуальных систем и структурного анализа НИУ ВШЭ, основатель нескольких стартапов в области ИИ, профессор ANECA, а также бывший сотрудник Oracle, представит доклад на тему “Дискурсивный анализ текста для организации диалога” на нашем традиционном Community Call на русском языке.

🤖Сделать диалог с чат-ботом логичным и интересным — важнейшая задача области Conversational AI. Для этого применяются самые разные подходы, и один из них — дискурсивный анализ текста. Его идея состоит в том, чтобы чат-бот помог пользователю сфокусироваться лишь на каком-либо предложении из всего текста. В дискурсивном дереве текст разбивается на части, связанные логическими отношениями, и чат-бот направляет по ним пользователя, развивая диалог. Например, это могут быть временные (temporal) отношения, когда пользователя наверняка заинтересует, что будет после описанного события или что было до него.

На нашем вебинаре Борис Галицкий подробно расскажет о способе управления ходом диалога в чат-боте на основе дискурсивного анализа текста.

❗Сохраняйте ссылку в ваш календарь, чтобы не пропустить Community Call

DeepPavlov

11 Jul 2022 17:39

📢 Завтра, 12 июля в 17:00 Research Product Manager в DeepPavlov.ai Мария Молчанова на научно-техническом вебинаре НТР проведет обзор методов и платформ для разработки диалоговых систем, а также расскажет о прогнозах их развития.

🤖Диалоговые системы плотно вошли в нашу жизнь: с ними мы взаимодействуем, когда хотим сделать заказ или же спросить у своего виртуального ассистента прогноз погоды. Среди диалоговых систем вам наверняка встречались чат-боты для сайта, приложений, мессенджеров, виртуальные голосовые помощники, персонажи в играх, роботы и т.д.

📑Мария в докладе подробно рассмотрит все эти виды диалоговых систем, а также методы их разработки. Помимо этого, она расскажет о наиболее частых проблемах и вызовах, с которыми сталкиваются разработчики различных типов диалоговых систем. И, наконец, продемонстрирует, как создать собственного multiskill ассистента на платформе DeepPavlov Dream.

☝🏻Вебинар открыт для слушателей, зарегистрироваться можно по ссылке.

DeepPavlov

01 Jul 2022 18:25

Hi, friends!
We have some amazing #ConvAInews for you in our traditional newsletter🔥

▪ Yesterday we had a rather inspiring Community Call with an invited guest Artem Rodichev. Daniel Kornev and Artem discussed the topic of the past, present and, of course, the future of Conversational AI. If you didn't manage to join the Call yesterday, you can watch a recording.

▪ DeepPavlov released the First Russian-language Distribution of the Dream Chatbot🤩Very soon you will learn all the details and will be able to chat with the bot.

▪ Tech giants Meta, Microsoft, Epic Games and 33 other technology companies have formed a group to standardize metaverse technologies. The Metaverse Standards Forum is designed to strengthen open, interoperable standards for augmented and virtual reality systems and 3D technologies.

▪ Amazon has announced the Alexa Prize Socialbot Grand Challenge 5 will open from July 6 to August 2023 for academic institutions interested in competing. The fifth Grand Challenge adds a new scientific invention and innovation award to the overall performance battle, with a total $1 million prize purse. The deadline for applications is October 5, 2022.

DeepPavlov

22 Jun 2022 12:49

На VC появилась статья о 5-ти самых популярных платформах для создания чатботов и виртуальных асисстентов - в их числе DeepPavlov Dream 🔥 Читайте и узнавайте про современные тренды развития и возможности ИИ-асисстентов по ссылке.

DeepPavlov

17 Jun 2022 10:36

Hello everyone!😀

Today, in our traditional newsletter we share with you not only #ConvAInews, but also the updates of the DeepPavlov team🎅

1️⃣ New Forums
▪️Now you can find the Releases and Tutorials sections for each product, which will be constantly updated.
▪️You can use the forum to ask any question about DeepPavlov products, share your user experience or ask the opinion of the community.

2️⃣ Playlist with workshops
DeepPavlov researchers have recorded video workshops that make it easier to create your own chatbots with the DeepPavlov Conversational AI stack. Check it out, leave comments and ask questions if something is unclear.

3️⃣ Rasa will stop supporting a free Community version of the Rasa X. The company says that they will keep it only for the Rasa Enterprise customers which have subscriptions.

4️⃣ Microsoft presented AdaTest, a process for adaptive testing and debugging of NLP models inspired by the test-debug cycle in traditional software engineering. It will offer an opportunity to combine the synergistic strengths of both user-driven approaches and automated approaches.

5️⃣ Amazon has named the University of Glasgow and its GRILLBot as the winner of the first Alexa Prize TaskBot Challenge. The TaskBot Challenge asked teams to produce a digital assistant that could help people with home repair, construction and culinary projects. GRILLBot and its competitors attempted to help users plan out their projects in terms of what they need to buy and how to combine the materials or ingredients into a successful meal or piece of furniture.

6️⃣ You probably heard the recent news about Google's LaMDA that has forced the continuous discussion on whether AI chatbots could be sentient. Blake Lemoine, a Google engineer, tested the chatbot and concluded that the chatbot has a mind. After reporting this in a document entitled “Is LaMDA sentient?”, however, he has been placed on leave. You can read the full story, including the dialogues with the system recorded by Lemoine, here. And you guys, do you believe that LaMDA is sentient?

DeepPavlov

15 Jun 2022 10:01

Друзья,

Мы приглашаем всех желающих на наш онлайн-семинар 17 июня в 17:00 (мск). Исследователи DeepPavlov Александр Чернявский и Олег Сериков расскажут про приложения мультимодального глубокого обучения в Conversational AI.

📑В своем докладе они расскажут про основные тренды развития обучения мультимодальных диалоговых моделей, трудности, возникающие при их обучении, в том числе про задачу data fusion.

🤖 Также будет сделан краткий обзор современных SotA среди аудиовизуальных моделей и методов, использующих совместно тексты и изображения, и полезных датасетов. Обсуждены приложения использования в сегменте русскоязычного рынка.

Присоединиться можно по ссылке.

DeepPavlov

30 Sep 2022 11:51

Hi, friends!
This is our regular #ConvAInews!

◾️ Daniel Kornev, CPO of DeepPavlov.ai, had the pleasure to be a guest on the Applied AI Podcast by Justin Grammens! Such an interesting talk about today and the future of Conversational AI! Follow the link to listen to the podcast.

◾️ Google AI Introduced ‘TensorStore,’ An Open-Source C++ And Python Library Designed For Reading And Writing Large Multi-Dimensional Arrays. Fundamental engineering problems in scientific computing linked to the management and processing of enormous datasets in neuroscience have already been resolved using Google’s TensorStore.

◾️ DeepMind’s new chatbot uses Google searches plus humans to give better answers. The lab trained a chatbot to learn from human feedback and search the internet for information to support its claims.

◾️ UK-based robotics company Engineered Arts just gave its ultra-realistic looking humanoid robot Ameca a voice. The company showed off Ameca having a conversation with a number of the company's engineers, a courtesy of a speech synthesizer and OpenAI's GPT 3, cutting-edge language model that uses deep learning to generate impressively human-like text.

◾️ OpenAI Starts Letting DALL-E 2 Users Edit Faces on Synthetic Images. People using OpenAI’s DALL-E 2 text-to-image engine can now upload and edit people’s faces.

Stay tuned!

DeepPavlov

13 Sep 2022 10:29

Hi, friends!

Daniel Kornev, CPO of DeepPavlov.ai, shared an opinion with AIbusiness.com on the topic: "Why open source is key to the future of Conversational AI".

Corporations could not pull something on the market until it's fully polished. Cooperation and joint efforts significantly drive innovation, and open source is the only way to develop great products and learn by making mistakes.

Do you want to learn more about other reasons why open source is crucial in ConvAi? Read the whole piece here.

DeepPavlov

08 Sep 2022 16:25

CV-Multimodality Engineer at DeepPavlov.ai

DeepPavlov is an open source conversational AI framework. We aim to help developers all over the world create chatbots, dialog systems, and Multiskill and assistants in a faster and easier way.

We are looking for an experienced researcher to work with computer vision models (CV) and models' architectures. The main task is to introduce the vision modality into dialog systems, and then work with a team building multimodal dialog systems.

If you have 2+ years of the experience with CV models, and the ability to write efficient, clean, readable code, you're the perfect candidate!

Here you can check details about the position:
https://deeppavlov.ai/careers/tpost/edgfl6htt1-cv-multimodality-engineer

#jobs

DeepPavlov

19 Aug 2022 08:13

Hello, everyone!

Today, Denis Kuznetsov, researcher at DeepPavlov.ai, will make a presentation on the topic: "Dialog Flow Framework: Overview".

He will talk about a young ambitious Dialog Flow Framework (DFF), an open source framework created for the development of modern dialog systems. DFF is easy to learn and is aimed at using machine learning models.

Denis will tell about DFF and cover the following topics:

〰️ Using DFF, it's easy to assemble the dialog assistant as a constructor using databases and channels through which messages come.
〰️ DFF has a set of tools for collecting and visualizing statistics of user behavior - for further analysis of the system and improvement.
〰️ DFF supports python as the main writing tool, but it also develops software for DFF in the manner of no-code\low-code development of dialog assistants.
〰️DFF is also actively developing solutions aimed at building a dialog assistant based on dialog data, with specified restrictions and with minimal human involvement in the development of a dialog system.

Date: 19 August (today!), 07:00 pm (GMT +3)
Here is the link on today’s seminar.

Please join us and learn more about simple and complex chatbots’ development with DeepPavlov.ai ☺️

#events

DeepPavlov

11 Aug 2022 10:53

Hello, everyone!

Daniel Kornev, CPO of DeepPavlov.ai, will be a speaker at the PyData conference in Yerevan. This event will be held this Friday.

Daniel will make a presentation on the topic: “Building your own Multiskill AI Assistant with DeepPavlov”.

Time: 15:30 pm (GMT +4)

Please join us and learn more about simple and complex chatbots’ development with DeepPavlov.ai ☺️

You can find out more details below:
https://yerevan2022.pydata.org/cfp/talk/39PYW3.ics

#events #conferences #convai

DeepPavlov

29 Jul 2022 16:32

Hey, friends!
Here's the #ConvAInews again!

▪ Yesterday we had a Community Call with Boris Galitsky, dedicated to the method of managing the dialogue in a chatbot based on discursive text analysis. If you didn't manage to get to our webinar, you can watch the recording on YouTube.

▪ New research by Luis Piloto and colleagues at Princeton University takes a step to make the computer compete with the common sense abilities of a typically developing infant. The researchers created a deep-learning artificial intelligence (AI) system that acquired an understanding of some common-sense laws of the physical world. More about it you can find here.

▪ Transformers achieve state-of-the-art performance for NLP, and are becoming popular for a myriad of other tasks. They are computationally expensive which has been a blocker to their widespread productionisation. How to use the BetterTransformer, read in the article by Pytorch.

▪ Amazon last week announced new tools and programs for third-party Alexa skill developers to grow their user base, increase retention, generate revenue, and keep more of what they earn. While Google is retreating from its third-party voice app developer ecosystem with the sunsetting of conversational actions for the Assistant, Amazon is extending its investment.

DeepPavlov

28 Jul 2022 13:53

Hello everybody!😊

Our new article about #DREAM Socialbot 🔥 has just came out!

Feel free to click on link, and find out more about Annotators!
We explained how different type of Annotators actually function and what problems they solve to make the dialogue more seamless and comprehensive.

DeepPavlov

27 Jul 2022 13:41

Друзья,
Напоминаем, что уже сегодня, 27 июля в 19:00 Диляра Баймурзина, Мария Молчанова и Лидия Остякова представят доклад об управлении диалогом. Подключиться можно по ссылке.

👨‍💻Одной из важнейших составляющих диалоговых систем является диалоговый менеджер - компонента, отвечающая за управление диалогом. Для каждого вида диалоговых систем существует несколько традиционных подходов.

🤖Например, задаче-ориентированные системы обычно используют фреймовый подход или подход на основе правил, которые используют различные аннотации моделей понимания естественного языка.

Однако на данный момент не существует общепринятого подхода к управлению диалогом с чат-ботами для общения на общие темы. Существующие виды управления диалогом:
·на правилах,
·вероятностные модели,
·гибридные алгоритмы,
·подходы на основе обучения с подкреплением -
Все подходы имеют свои преимущества и недостатки.

Важной особенностью большинства подходов является отсутствие многошагового планирования диалога, поэтому также в докладе исследователи DeepPavlov рассмотрят теоретические аспекты управления диалогом на основе анализа дискурса и определения целей пользователя и системы в диалоге.

DeepPavlov

21 Jul 2022 15:14

🤖Количество желающих стать частью области Conversational AI растет с каждым днем. И такая возможность становится все доступнее благодаря различным программам стажировок.

У нас в DeepPavlov 🎅 тоже есть стажировки, где будущий специалист может выбрать направление в сфере разговорного ИИ, которое ему по душе.

Больше о стажировках в разных организациях читайте здесь.

DeepPavlov

15 Jul 2022 12:05

Hello, friends!👋🏻
We’ve got a fresh newsletter with #ConvAInews for you!

This Week’s Interesting Picks:

▪ We remind you that on July 27 at 19:00 we will hold a public seminar in Russian where Dilyara Baymurzina and Maria Molchanova will present a report on dialogue management. The researchers will talk about various ways to manage the dialogue with Open-domain chatbots. Join via this link.

▪ The Alexa Live 2022 conference will be held on July 20. Amazon’s annual Alexa developer conference is dedicated to the future of ambient intelligence, smart home, and voice AI. You can register by following the link.

▪ Meta has made the NLB-200 model publicly available for translating text into 200 languages. The "No Language Left Behind" project is part of Meta's plans to support rarely used languages and develop a universal translator. Meta will use NLB-200 to improve translation on Facebook, Instagram and eventually in the metaverse.

▪ Apple discontinued the HomePod over a year ago, but the smart speaker may make a comeback next year, according to a Bloomberg report. The revamped HomePod will supposedly be similar in size to its predecessor rather than the smaller HomePod Mini currently available and will use the same S8 processor that will be in the upcoming Apple Watch. The report also suggested a new Apple TV and new AirPods Pro earbuds are on the way from Apple.

DeepPavlov

08 Jul 2022 14:44

❗️Напоминаем, что сегодня в 19:00 можно подключиться на наш семинар по обзору современных подходов к определению эмоций в диалоге от исследователя DeepPavlov Яны Шишкиной

DeepPavlov

30 Jun 2022 17:30

Hello, friends!😊

We remind you that our DeepPavlov Community Call about The Past, Present and Future of Conversational AI with a special guest Artem Rodichev, Founder and CEO of Ex-human & Ex-head of the AI at Replika, and Daniel Kornev, CPO at DeepPavlov, is starting in 30 minutes!🔥

📌Join us via this link

DeepPavlov

20 Jun 2022 14:16

❗Друзья,

В прошлую пятницу у нас уже состоялся публичный онлайн-семинар по мультимодальности. Этим летом у нас подготовлено еще 4 доклада по всем направлениям исследований DeepPavlov.

👨‍💻Мы приглашаем всех интересующихся областью ИИ подключиться и узнать больше о наших направлениях деятельности.

Расписание:

• 29 июня, 19:00 — Дмитрий Евсеев расскажет об использовании баз знаний для ответов на вопросы и генерации реплик в диалоге.

• 8 июля, 19:00 — Яна Шишкина на семинаре проведёт обзор современных подходов к определению эмоций в диалоге, генерации эмоционально окрашенных ответов и тех инструментов и ресурсов, которые для этого используются.

• 27 июля, 19:00 — Диляра Баймурзина и Мария Молчанова представят доклад об управлении диалогом.

• 19 августа, 19:00 — Денис Кузнецов разберет в докладе последние тенденции в построении диалоговых ассистентов на примере решения с открытым исходным кодом Dialog Flow Framework .

📌 Для всех семинаров единая ссылка

DeepPavlov

16 Jun 2022 11:11

Dear DeepPavlov community,

🔥On June 30, a special guest Artem Rodichev, founder and CEO of Ex-human & ex-head of the AI at Replika, and Daniel Kornev, CPO at DeepPavlov, will hold a discussion about The Past, Present, and Future of Conversational AI.

🤖Socialbots are getting into our lives more and more: they are not only assistants who fulfill our requests, but also means to communicate. So, Artem and Daniel will describe several SOTA chatbot models which can generate very impressive examples of human-model conversations. What are Digital Humans? How do the Task-oriented and Open-domain chatbots fuse? What will the world look like in 2030? You will find out the answer to all these questions on our Community Call.

❗Save the link to our call and don’t forget to add it to your calendar.

DeepPavlov

13 Jun 2022 13:37

Завтра, 14 июня, будет еще одно выступление на конференции от DeepPavlov 🔥

Мария Молчанова, наш Research Product Manager, представит доклад о современных методах для построения сложных диалоговых систем на конференции Data Start 2022 🤩

👨‍💻Создание диалоговой системы, которая сможет поддерживать глубокий и осмысленный диалог на различные темы - одна из важных и сложных проблем в области искусственного интеллекта. При этом разные компании уже предлагают свои решения для разработки собственных ИИ-ассистентов как Task-Oriented и Closed-Domain, которые работают по четкому алгоритму, так и Open-Domain, способных пообщаться на любую тему, заданную пользователем.

Так, в своем докладе Мария осветит следующие ключевые аспекты:

• Проблемы и задачи при создании диалоговых систем
• Обзор решений для создания ИИ-ассистентов
• Основные подходы при разработке Task-Oriented и Closed-Domain систем
• Основные подходы при разработке Open-Domain диалоговых систем
• Разработка ИИ-ассистента на примере платформы Dream

❗Зарегистрироваться можно как на бесплатный трек, так и на более расширенный платный по ссылке.