Category: Tech info

A web service solution or platform for task automation, IFTTT is a service that interconnects hundreds of platforms, websites and objects to allow users to automate their most common actions.

IFTTT: what is it?

Launched in 2010, IFTTT is a service that allows you to connect devices to each other and automate certain tasks according to personalized scenarios. In short, IFTTT is a service that bases its operation on programming to automatically trigger an action as soon as another first action occurs. Thanks to IFTTT, it is possible, for example, to trigger your heating as soon as the temperature drops below a certain threshold. Accessible on the web, Android and iOS, IFTTT offers a list of applets (previously called “recipes”) representing all existing actions, which can be configured according to needs.

Is IFTTT free?

The IFTTT web service went live in September 2020 with a paid subscription. By the end of September 2022, the service had 150,000 users, generating $6 million in annual revenue.

Until now, the platform was free and IFTTT users could create as many applets as they wanted even if they were limited to one trigger. Now, this free service is limited to five applets.

Subscribers to the Pro license, sold at 2.6 dollars per month, have access to 20 applets. You have to pay 5.2 dollars per month to access the Pro+ license, which allows you to create an unlimited number of applets and benefit from customer support.

Why use IFTTT?

IFTTT can be used in many professional fields to automate repetitive tasks for which human intervention is not essential. For the general public, IFTTT can be used to live in a connected home with, for example, the creation of a recipe that turns off the lights as soon as a person leaves the house.

Which apps and web apps are supported by IFTTT?

IFTTT currently supports many apps including the must-have Gmail, Facebook, Instagram, WhatsApp, etc. In the case of these apps, IFTTT can allow, for example, to create an applet (or recipe) in order to receive an email as soon as someone posts on Twitter with an identified hashtag, to turn on a light when the Uber Eats delivery driver is ready to deliver, etc.

End of IFTTT on Google

As of August 31, 2019, several Nest products are no longer supported by IFTTT. Google now favors the control of its objects with Google Assistant. For the web giant, this decision simplifies the use of its products.

How to use the IFTTT API?

The IFTTT API is used like most web services dedicated to task automation. Compared to the Zapier API, the IFTTT API is less complex to learn.

IFTTT vs Zapier

IFTTT offers much the same functionality as another task automation platform, Zapier created in 2011. However, there are some differences between these two platforms, especially in terms of availability as apps for iOS and Android and possibilities as to the integration of business applications.

What are the alternatives to IFTTT?

Besides Zapier, there are platforms or services developed by Google (Google Assistant: Actions on Google), Microsoft (Microsoft Flow) or Amazon (Alexa Skills Kit).

Most users of the free version are content to manage the existing applets on IFTTT that meet their needs. But they are limited by the fact that only one action is possible between two services, and only two services can be connected together. As a result, for an intensive or experienced user, it can quickly become tedious since it is necessary to initially create and parameterize multiple applets to design a chain of consecutive interactions mixing different services. With IFTTT Maker, this problem has been solved: it is possible to cumulate several actions in a chain, by adding as many instructions as necessary. IFTTT Maker also allows designers, developers and companies to add new services to make them accessible to the greatest number of people. This is particularly useful for connecting the new connected objects that arrive on the market every year.

Originating from the Japanese automotive industry, Kanban aims at the continuous improvement of processes. Companies massively use this agile method and its flow visualization mode in the form of a board for their project management.

What is Kanban?

Kanban is a work method inspired by the Lean approach focused on the continuous improvement of production processes. It was conceptualized in 1950 by the Japanese industrial engineer Taiichi Ōno for Toyota in order to optimize the car manufacturer’s manufacturing. It is based on a “pull flow” system, taking into account consumer demands, and not push flow.

Kanban thus consists in constantly adapting to the customer’s needs. Objective: to limit the risk of overproduction and waste, but also to reduce delays and costs. Qualified as an agile method in the same way as Scrum, Kanban advocates the visualization of workflows through a Kanban board. A board allowing to prioritize and follow the progress of tasks to be accomplished.

Why use a Kanban ?

A Kanban board is a project management method that aims to optimize team work through a visual modeling of tasks and workflow. The challenge of this method is to limit the number of tasks in progress at each stage of the project to make the work flow more smoothly and efficiently. These optimization actions will be carried out during regular meetings of the project team, with the objective of rationalizing the distribution of tasks over time.

The Kanban method is based on a card system. It was named after this process. Kanban means labels in Japanese. The cards or labels will represent the tasks to be accomplished in order to respond to a customer’s order.

These graphic objects are a simple tool to understand and quickly decipher. To visualize the progress of an order or a project, they will be distributed in a Kanban board generally divided into three parts: “To do”, “In progress” and “Done”. This way, everyone knows what they have to do, when and how.

The creation of a kanban board for a given project involves several steps:

1. Identify the successions of tasks to be performed that can be modeled as collaborative flows,
2. Map each phase of the process,
3. Identify phases that are resource intensive or dependent on many subtasks,
4. Name the steps/columns in the table according to the business workflow.

The definition of the columns or lists in the table depends on the business process. For example, a content publication workflow could be broken down into the columns “Proposed topics”, “Validated topics”, “Drafting in progress”, “Completed articles”, “Reviewed articles” and “Published articles”. Obviously, the stages will be different for an IT project, which could be broken down into “To be developed”, “Under development”, “Completed development”, “Testing phase”, “Acceptance phase”, “Deployment”.

What are the 3 types of Kanban?

A Kanban card will be able to contain three types of information:

1. the description of the task to be accomplished,
2. The contents useful for the realization of the task: reference links, drafts, models,
3. Comments submitted by project team members,
4. The subtasks to be completed to accomplish the task.

What is the role of kanban daily meetings

Agile methods have revolutionized traditional meetings around a table. We now use daily meetings. These are short meetings (no more than 15 minutes) that are held every morning, usually standing up. Each team member must answer three questions: What did I do yesterday? What will I do today? What are the obstacles I am likely to encounter?

Daily meetings allow the whole team to be informed of each other’s tasks and to find solutions to certain problems. As a result, they are often considered as a complementary tool to the Kanba approach.

Kanban with Jira

To implement the Kanban method, many companies use the Jira software. Designed by the Australian editor Atlassian, this application implements the visualization approaches promoted by the Scrum and Kanban agile methods. It allows to create customizable and interactive Kanban boards.

Kanban with Trello

Directly inspired by the Kanban methodology, the Trello application is defined as a project management tool in SaaS mode. This service offers a free version and a paid version with more advanced features (especially in terms of card management, automation and security).

One of Trello’s main assets is its collaborative work features. In particular, it allows members of a project team to post their comments in the tags.

Kanban vs Scrum

Scrum advocates short development cycles (or sprint), from two to four weeks, at the end of which a new version of the project is delivered. As we have seen, Kanban emphasizes continuous improvement, with permanent evolution possible.

The two approaches therefore seem to be in opposition. They are nonetheless complementary, and are often used together. A kanban board can indeed be used to manage the progress of tasks to be completed within a Scrum sprint.

A feature store manages the storage, updating and sharing of features of machine learning pipelines and models. A solution that is essential for an AI factory.

What is the feature store?

The notion of feature store was first introduced by Uber in a 2017 published post. Under this term, the VTC refers to the repository it uses to store the features of its machine learning models. The challenge? Making it easier to share these features from project to project. Uber claims to be confronted with numerous modeling problems involving similar or identical attributes. Hence the idea of pooling them through a central database.

In the context of an AI factory, the feature store is central. It allows to capitalize on the learning functions already created for new developments. So much time is saved in feature engineering. In the field of health, features can refer to an individual’s blood type, height curve, weight… The feature store centralizes the features and allows them to be easily shared with other machine learning models.

Example of a feature store

Features refer to the information used to feed the machine learning. In the case of a recommendation AI on a music streaming platform, it is for example the songs already listened to, their playback time or their ranking in terms of audience.

Thanks to a feature store tool, the streaming platform will typically be able to reuse the traffic ranking used here for other models, on the advertising targeting front for example.

What is the advantage of a feature store?

The feature store consolidates machine learning workflows around a single pipeline for training, testing and validating models, as well as executing them in the field. It allows you to benefit from a single source of truth with a single data transformation method per feature. Federating transformation methods in this way facilitates their monitoring and validation, and simplifies the tracking of biases that may arise from the features or data. The repository also stores feature metadata and their history. This makes it possible to keep track of the comments made by data scientists on the influence of features on a given model, and thus to better identify the types of problems to which they can respond.

During the training phase, the feature store is also there to guarantee the integrity of the data sets. “The training data may mistakenly include information dedicated to testing the model once it has been trained. This obviously distorts the results. This is called a feature leak,” explains Ismaïl Lachheb at Octo Technology. Such a leak happens quickly. “In the case of a database with a dozen joins (between the tables, editor’s note), all it takes is an error in one of them for the learning set to access data intended for testing,” says the data scientist. During the creation of the various data sets, the feature store is responsible for guaranteeing the watertightness of the data, regardless of the number of joins and tables used. “It manages the versioning and execution of the learning process in line with the evolution of the state of the data over time,” adds Sergio Winter, machine learning engineer at Revolve, a Devoteam entity with expertise in AWS.

The last major benefit of the feature repository is that the feature store ensures standardization of feature formatting and calculation between learning and real-world prediction. “If the data preprocessing is not exactly the same in both cases, a learning/invocation bias will appear with a negative impact on the quality of the results,” warns Ismaïl Lachheb. This difference may be due to negligence or to the management of the learning and inference data sources by different teams. By controlling the featuring of the data during both prediction and training, the feature repository ensures consistency between the two sources.

What are the feature store tools?

To get started, there are several open source feature store tools. The most popular ones are :

• Feast,
• Hopsworks,
• Tecton.

Tecton was created by the developers behind Uber’s AI platform (read the article Feature store comparison: Tecton outshines Feast and Hopsworks).On the cloud side, Amazon Web Services (AWS) and Google also market managed feature repositories. AWS’ offering has the advantage of integrating with Data Wrangler. “Unlike Google’s Cloud Dataprep (which is based on a third-party application from Trifacta, ed.), it’s a graphical tool that not only handles batch but also data transformations and data set updates in real time,” compares Sergio Winter (see the article AI Cloud Platforms: Amazon and Microsoft outpaced by Google).

Databricks Feature Store

Databricks offers its own AI platform: Databricks Machine Learning. Designed to run on the editor’s infrastructure designed to federate big data analytics and machine learning, it includes a feature store.

Feast: open source feature store

Feast is an open source feature management library. This library allows to define feature stores to help building models and retrieving data.

Feature store vs data warehouse

A data warehouse is a data warehouse (or EDD). It is a relational database that collects data from a wide variety of sources. Its main function is to validate an analysis and optimize the decision-making process of a company.

The feature store is a kind of data warehouse (feature-oriented) at the service of machine learning. The feature store is different from an architectural point of view, insofar as it is a double database with its own particularities:

• A database that contains data distributed by the SDK (Software Development Kit), with a large temporal depth.
• A database that contains fresh data and streaming data, this DB (database) is faster to serve “fresh” data.

Narrow artificial intelligence or Narrow AI, is AI focused on a given task. Its machine learning model is trained to solve a specific problem.

What is narrow AI?

Narrow artificial intelligence (or Narrow AI) focuses on a specific problem, in this sense it is monotask. narrow AI aims at reproducing specific cognitive faculties, such as natural language processing or image recognition for example.

We can thus say that it simulates a human behavior or faculty, but without consciousness. It is a particularly powerful tool in the sense that it is able to automate tasks. It draws from the algorithms of machine learning.

What about some examples of narrow artificial intelligence?

The current applications of narrow artificial intelligence are numerous. narrow AI allows indeed to treat precise tasks, and in a faster way. In short, to be as good or even better than a human brain when performing these same tasks. Among the application cases of narrow AI, we can mention :

• Understanding and interpreting (to some extent) natural language,
• Reproduce the musical style of a famous musician,
• Produce images that appear to be grounded in reality,
• Recognize and identify people in a photograph,
• Make a diagnosis of a disease.

Siri, the intelligent assistant on Apple devices, is a perfect example of narrow artificial intelligence. Able to understand certain requests, it answers a number of questions asked by humans. It is thus able to understand and interpret language, or part of it. This is a technological feat that has revolutionized usage. However, it has only a small number of predefined functions and there is neither true intelligence nor consciousness.

What can narrow AI do?

While applying narrow artificial intelligence, professions could be automated. Some studies show that in the context of the coming decades, 10% of jobs could be eliminated, and 50% transformed due to the potential intervention of narrow AI in certain professions.

What is the difference between narrow and strong AI?

Narrow artificial intelligence reproduces a human mode of operation without consciousness or sensitivity. There is no “true” intelligence.

The ambition of strong artificial intelligence is to develop models allowing the creation of machines endowed with mind, consciousness and sensitivity. These machines would have the subtlety of the human brain coupled with the performance and computing power of AI as it exists today. This is one of the main challenges for artificial intelligence.

The Notion platform federates task and project management, note taking, online document editor… Oriented no code, it is presented as an all-in-one digital workplace oriented to team management.

What is Notion?

Claiming 20 million users worldwide, Notion.so is an all-in-one digital workplace. It combines various collaborative tools for note taking, task management, project management (around a kanban board) and document storage and exchange. Oriented to no-code development, its graphical interface, in wiki mode, is fully customizable. Like Airtable, it allows you to create a database. But also formulas to automate calculations, display content under certain conditions, or edit and share web pages.

The icing on the cake, Notion is equipped with an API designed to connect the platform to third-party applications: Office 365, Google Workspace, Slack, Salesforce … According to the latest Okta Business at Work 2022 barometer, Notion is the fastest growing app in terms of adoption in 2021, both in terms of number of customers (+240%) and users (+663%).

If you are not familiar with task management applications, Notion.so may take some getting used to. The way it works is quite simple. Each task or note is synchronized with different views: document, calendar, timeline… It is also possible to create a mini-database through a data table, and to apply tags, filters, sorting or calculate sums and percentages.

Available online, access to Notion involves creating an account by filling in a simple e-mail address. The solution can then be used immediately at no cost. Called Teams, a feature currently being deployed will allow to organize documents according to distinct teams (development, support, marketing, finance…) or projects (product launch, conference organization…).

Is Notion free?

Notion.so offers a free plan, called Notion Personal. It targets small teams of up to 6 people. Allowing to create an unlimited number of pages and blocks, Notion Personal opens access to the platform’s API. Besides its web version, it allows to use Notion via its different applications (for Windows, Mac, Android and iOS).

Depending on your needs, several paid subscriptions allow you to unlock these limitations while giving access to more advanced features (see below)

What is the price of Notion.so?

Notion has three paid offers. Notion Personnal Pro shows a price of 4 dollars per month, Notion Team 8 dollars per month per user. As for Notion Enterprise, its price is not public.

Notion Personnal Pro is limited to the management of a single team, but gives the possibility to invite an unlimited number of participants. Unlike the free offer, it manages a version history (over 30 days) and an unlimited volume of files. Unlike Notion Personnal Pro, Notion Team is designed for multi-team management. It introduces administration tools, as well as collaboration workspace to control access rights individually or by group. Finally, Notion Enterprise extends the functional palette to include support for single sign-on (via SAML), user management via the API and more advanced access policy management.

In addition to its website, Notion is also available as applications for Windows, Mac, Android and iOS:

Notion templates

Notion offers a gallery with several hundreds of templates, most of them developed by external contributors. This Notion Template Gallery federates very diverse application templates: weekly agenda, daily planning, to-do list, kanban board, budget management, roadmap, company journal, mini-CRM, recruitment pipeline… It includes templates for both personal and collective organization.

Notion API

Accessible from the free mode, the Notion API allows to automatically import data from third party tools or conversely to manage exports. And this with about twenty applications: Asana, Confluence, Trello, Office 365, Google Workspace, Slack… Another possibility offered by the programming interface is to trigger an update in a third-party software in response to an action performed in Notion (change in the status of a task, a document, etc.), or to send an email, or an alert by SMS, etc.

For the future, Notion is working on the Synced Databases feature which will allow to synchronize its database with structured data from third party tools like Google Calendar, GitHub or Jira.

Does Notion.so integrate with Zapier or IFTTT?

Notion.so benefits from the task automation tools Zapier and IFTTT. The first one allows to connect the platform to hundreds of applications, and the second one to about twenty. The advantage of these two tools? To offer a no-code integration environment that is much easier to use than Notion’s API, which by definition involves getting your hands into the code.

Is Notion.so usable offline?

Notion.so offers an offline use of its services on any type of device, including in web mode. This is a particularly practical feature, especially when traveling.

Is Notion.so available on Android?

In order to offer a smooth and pleasant experience, Notion.so can be used on Android and iOS from a smartphone. The editor offers applications for both mobile systems, which provide access to the same features.

Notion : tutorial and other links

Since mastering Notion can be a bit tricky as the possibilities are so extensive, using tutorials and forums can be a good idea.

What are the alternatives to Notion?

Comparable solutions to Notion.so include clickup and monday.com. To date, these are without a doubt its main competitors. The first one is comparable in terms of features, templates and price. As for Monday, it has a plethora of templates, but its prices are much higher. Its entry price is 8 dollars per user per month.

Complementary to agile methods, design thinking is a creative design approach based on brainstorming and collaborative creation.

What is design thinking?

Design thinking is a collective innovation approach. At the crossroads of analytical and intuitive thinking, it recommends involving future users in a collaborative creativity logic. The idea is to get them to capitalize on collective intelligence to come up with innovative product or service prototypes.

If design thinking has come back into fashion in recent years, it is a method that originated in the 1950s with the invention of brainstorming by the American advertising executive Alex Osborn, a concept that made the business world aware of creative thinking. The concept of design thinking was more widely evoked in the book “Visual Thinking” by Robert H. McKim, published in 1973. It was then developed in detail in the 1980s by Rolf Faste at Stanford University.

What is the purpose of design thinking?

The aim of design thinking is to stimulate innovation in an organization, whatever its size, by organizing collective creation sessions. The objective is to come up with a concept or even a prototype of an innovative product or service.

Design thinking is based on four main principles:

1. Identify the problem to be solved,
2. Take into account the users’ needs,
3. Fostering collective and multidisciplinary intelligence,
4. Test several solutions, do not be afraid of failure to progress.
   

What are the three basic criteria of design thinking?

Design thinking takes shape in a process that was defined by the American designer Rolf Faste. It is based on three major complementary evaluation criteria:

1. Desirability which refers to the question of whether users really want the proposed solution, and ultimately whether they will adopt it,
2. Feasibility which implies having the necessary resources to develop this solution (budget, technologies, skills…).
3. Sustainability which echoes the interest of deploying such a solution, and more precisely its profitability, i.e. its potential return on investment.

What are the 5 phases of design thinking?

According to the Stanford University d.school, design thinking can be reduced to 5 stages which are:

1. Empathy With customers/users to get to know them,
2. The definition of the problem to identify their expectations,
3. Idea generation to find an answer to these expectations,
4. Prototyping to shape the solution,
5. The test phase to test the final product.

What are the main design thinking tools?

The design thinking method relies on many tools, such as the customer journey map which traces all the interactions of a consumer with a product, to create a user journey. It is also possible to use affinity diagrams, moodboards, personas or mind maps. Then come the tools specific to design:

• Brainstorming,
• Co-creation meetings,
• Idea menus,
• The positioning matrix.

Some design thinking use cases (examples)

Combining the analytical thinking of engineers or marketers and the intuitive thinking of creatives, design thinking focuses on the user experience (UX). Going through three key stages (inspiration, imagination and implementation), design thinking can be applied to all areas involving interaction between the user and the product or service. From there, design thinking can be found in many fields: fashion, software man-machine interfaces, car interiors and dashboards, interior design, smart cities, connected objects…

Several famous brands have used design thinking to find innovative solutions for their customers. For example, Lapeyre called on students from d.school Paris. Their objective was to rethink the bathroom to adapt it to seniors. After more than a year of discussions with senior citizens, caregivers and carers, they identified the needs of the elderly: to be able to sit down, to protect their belongings from water, to be able to see better up close, to be able to move around easily, to have good lighting, etc. By analyzing these needs and implementing the 5 steps of design thinking, they were able to create the “Concept” Care modular furniture.

With the design thinking method, personas are used in the definition phase. They allow to make visible and concrete the users to better understand their expectations. Personas allow marketing teams to remember who their target is when developing a product or service.

How do design thinking and agility complement each other?

Agile methods and design thinking are complementary, not to say intimately linked. In the agile project management cycle, design thinking is positioned in the solution research phase. The various agile methods then cover the way in which this solution is developed by implementing regular feedback between the team of developers and the end users. Their objective: to create and deploy a product through successive iterations, constantly validated by the business client, both upstream and downstream, in order to avoid any drift.

Design thinking training

Taking up all these theoretical writings, design thinking training teaches this entrepreneurial approach which is now most often summarized in three key steps: identify a problem and understand its environment, find the idea that will solve this problem, and transform the idea into a concrete project.

The teaching of design thinking insists on the fact that this entrepreneurial method requires a logic of co-creation, intellectual gymnastics between analysis and intuition, but also a field study. If design thinking can be learned through workshops or Mooc. IBM makes its internal design thinking guide available to everyone online.

Alexa, Google Home and Siri voice assistants are all based on automatic language processing technologies. Objective: to have the ability to understand, process and generate voice messages.

What is natural language processing (NLP)?

Natural language processing (NLP) is a branch of artificial intelligence that focuses on giving machines the ability to understand, generate or translate human language as it is written and/or spoken. Chatbots are among the most popular NLP software. Other famous NLP applications are the voice assistants Alexa, Google Home or Siri.

Why NLP?

Natural language processing aims at providing software with automatic processing of vocal or textual language. From there, it covers many cases of use more or less elaborated:

• Text classification,
• Text recognition,
• Automatic summarization,
• The automatic translation,
• Chatbot, voicebot or callbot,
• Intelligent assistants.

NLP combines artificial intelligence and language processing. The latest generation of NLP technologies is based on artificial neural networks or simple statistical machine learning models. Learning models will have been trained on large volumes of text.

The objective can target several types of automatic processing: speech-to-text and text-to-speech, named entity recognition (names of people, places…), sentiment analysis (positive, negative, neutral), text summarization, aspect extraction (targeting the intent of the text) or topic modeling.

What are the two types of machine learning models in NLP?

Overall, natural language processing falls into two broad categories of machine learning models:

• The machine learning models oriented NLU (natural language understanding), which are concerned with grasping the meaning of a language and a discourse in its context,
• The machine learning models oriented NLG (natural language generation) which aim to generate a text in the manner of a human.

Some examples of NLP oriented deep learning models

Although natural language processing has been around for a long time, recent progress in this field has been considerable. The number of NLP projects oriented towards artificial intelligence has multiplied among the digital giants. Among the most advanced models, we can mention:

• Google AI’s BERT and ALBERT models,
• Models derived from this first family such as RoBERTa (Facebook), StructBERT (Alibaba), DeBERTa (Microsoft), DistilBERT (Hugging Face),
• Alternative models such as GPT-2 and GPT-3 (OpenAI), XLNet (Carnegie Mellon University), UniLM (Microsoft), and Reformer (Google).

With the development of artificial intelligence, robotics is booming. Automata are now joined by humanoid robots or autonomous computer programs: bot, chatbot…

What is robotics?

Robotics is a field of activity covering the study, design and manufacture of robots or automated machines. It involves skills in computer science, electronics and mechanics.

What is a robot ?

To be considered as a robot, a machine must be equipped with sensors, which analyze the environment, but also with a logic processing system and actuators.

From the information obtained in its environment, the latest generation robot can act autonomously and intelligently based on embedded learning models or machine learning. It thus enters the AI family.

What is the origin of the word robot?

The term robot was conceptualized in 1920 by the Czech writer Karel Çapek. He used this term in one of his plays (Rossum’s Universal Robots) to designate an android invented by a scientist to perform tasks normally done by humans.

Robots types:

Robotics encompasses a very broad and diverse field of activity. Here are some examples and fields of application:

• Industry: assembly line management robot…
• Army: drone, spy robot, mule robot…
• Security: video surveillance…
• Health: ultrasound, assisted surgery…
• Aerospace: NASA explorer robot…
• Transportation: autonomous car…
• Domestic use: vacuum cleaner robot, lawnmower robot…
• Accompaniment: automated toy, humanoid robot…
• Computer science: chatbot, voice assistant…

Robots revolution ?

1. The first robots were nothing more than mechanical automatons. In the 16th century, Leonardo da Vinci is said to have created the first android automaton that could coordinate the movements of its limbs and jaws.
2. Cybernetic robots emerged in the early 20th century. They are robots programmed to perform specific tasks. The most advanced are the first rovers sent to other planets.
3. Humanoid robots have emerged from the 2010s. These are robots that take on a human appearance and are designed to interact with humans. This generation integrates the latest developments in AI, especially in deep learning.

What are the three laws of robotics?

In order to limit the behaviors or harmful effects that could be caused by robots with autonomy, three laws of robotics have been enacted. These laws were enacted by the science-fiction writer Isaac Asimov and the editor John Campbell. These three laws appeared in the short story “Runaround” published in 1942:

• A robot may not harm a human being or, by remaining passive, allow a human being to be exposed to danger.
• A robot shall obey orders given to it by a human being, except where such orders conflict with the first law.
• A robot must protect its existence so long as such protection does not conflict with the first or second law.

What are the related jobs in robotics?

Among the jobs in robotics are robot designer, maintenance technician, mechatronics technician (a profile combining electronic and computer skills) or digital control operator.

When did robotics first appear?

The term robotics, derived from the word robot, was first used by the Russian writer Isaac Asimov. It is found in the science fiction story “Liar!”, published in 1941.

Robots, inspired by the first automatons, appeared at the beginning of the 20th century. They are equipped with sensors allowing them to collect information on their environment, which then influence their activity. The development of robotics is strongly linked to the industrial field, which has developed since the middle of the 20th century. The latest intelligent robots created are autonomous objects, voice assistants, humanoid robots…

Speech synthesis allows the machine to translate the voice into text. With artificial intelligence, its operation becomes optimal. Many free speech-to-text tools exist.

What is speech synthesis?

Speech synthesis is a computer technique that consists in generating an artificial voice. It relies on linguistic processing techniques to convert the text produced by the machine into a phonetic version, and then on signal processing techniques to transform the latter into a digital sound rendering that can be broadcast through a loudspeaker.

In contrast to speech-to-text (automatic speech recognition), text-to-speech refers to the transformation of computer text data into artificial voice.

What is a synthesized voice?

A synthesized voice results from the conversion of a text into a sequence of phonemes aiming to be as close as possible to a human voice.

What is the best text-to-speech?

In 2020, the Mozilla Foundation published a study in collaboration with Carnegie Mellon and Northwestern Universities to estimate the quality of text-to-speech applications. Google’s wavenet text-to-speech model came out on top of this benchmark, ahead of Windows and Amazon Poly.

Free text-to-speech tool

There are many free text-to-speech tools available online. They are designed to translate text into voice on the fly. Some of these voice generators include:

Text-to-speech in Word is offered in Office 2019, Office 2021 and Microsoft 365. To activate it, you need to follow these steps: from Azure Text to speech for the first

1. Go to the Review menu,
2. Click on “Read Aloud”,
3. From the command menu, select “Read” to have Word read the text aloud.

Realistic speech synthesis

Amazon, Google and Microsoft each offer text-to-speech cloud services based on giant artificial intelligence engines. The goal: to get the most realistic voice possible. The first one is Amazon Polly, the second one is Google Cloud Text-to-Speech and the third one is Azure Text-to-Speech.

To create a synthesized voice, it is necessary to use a text-to-speech (TTS) engine. It allows to shape autonomously an artificial voice from a text produced by the machine, thanks to the use of AI and deep learning.

It is important to differentiate between intelligent TTS software and simple automated voice response software, based on pre-recorded words in a database. There are many websites offering a free text-to-speech service (see list above).

What is a text-to-speech engine?

A text-to-speech engine consists of a front-end and a back-end. The front-end is dedicated to splitting the text into words by associating each one with its phonetic transcription. This phonetic analysis step by fragmentation of the text occurs first.

Then, the voice synthesis engine uses its back-end system (synthesizer) to convert the linguistic and phonetic strings thus cut into sound. The synthetic voice is created by this last process.

What is the contribution of deep learning in speech synthesis?

Deep learning, through the use of deep artificial neural networks, allows to optimize speech synthesis in order to bring the sound rendering closer to the human voice. It reproduces voice inflections, intonations, tone variations, even accents…

Deep learning also introduces changes in rhythm and pronunciation. These elements will contribute to a better understanding by the target audience, but also more flexibility in language programming.

On Android, Google offers a text-to-speech application. Available on Google Play, it allows to use an Android smartphone to control applications by voice. In concrete terms, it translates a vocal request into a written request that can be understood by the software. It also allows to transcribe a text that has been recorded vocally.

Within its cloud offer, Google proposes a speech-to-text API for developers. It allows them to integrate text-to-speech features into their applications on a pay-per-use basis.

Examples of speech-to-text applications

Speech synthesis concerns different fields of application, such as:

• Audio books,
• Audio working documents (example of ReadSpeaker software that also underlines the written text),
• Techniques for reading without looking at a screen (visually impaired),
• Intelligent automated telephone services,
• GPS,
• Bank machines with integrated voice,
• Voicebots,
• Intelligent voice assistants (Alexa, Google Home…)…

Reinforcement learning reproduces the “natural” mechanism of knowledge acquisition. Robot, chatbot, autonomous car, its applications are multiple in artificial intelligence.

Reinforcement learning implements learning algorithms that learn from repeated experiences by trial and error. It thus reproduces the “natural” mechanism of knowledge acquisition.

To guide learning in the desired direction, reinforcement learning algorithms validate the decisions made by the machine via a reward or penalty mechanism. One could compare the process to training.

What are reinforcement learning algorithms?

The two most famous reinforcement learning algorithms are TD learning (for temporal difference learning) and Q-learning. These learning models are inspired by the human (and animal) process of acquiring knowledge through trial and error.

What is the advantage of reinforcement learning?

The main advantage of reinforcement learning is that to program a robot, for example, there is no longer any need for long and tedious development work. The computer will learn to operate, to react to this or that event or request by itself.

Whether the robot is physical or virtual, the learning phase will be carried out in the form of a digital simulation. This operating mode optimizes the learning time.

What is deep reinforcement learning?

Classified as automatic (or unsupervised) machine learning, reinforcement learning generally relies on neural networks in order to efficiently estimate the validity of a “complex” strategy, with a large number of choice criteria to be taken into account. This is called deep reinforcement learning (DRL). The main challenge is to achieve a system that encourages the desired behaviors, without undesirable side effects.

Often presented as the ultimate AI, DRL allows the creation of software capable of reaching or even surpassing human intelligence in several domains. The most famous system taking advantage of the method is none other than DeepMind, Google’s AI platform (resulting from the 2014 acquisition of the British company of the same name). It is on it that the American giant based itself to develop AlphaGo, the supercomputer known for having defeated in 2017 the world champion of go, the Chinese Ke Jie.

Examples of reinforcement learning

Deep reinforcement learning is used in many areas:

• Robotics in factories and warehouses to allow automata to learn by themselves how to lay down a new part model without prior programming.
• Calibration and quality control of industrial systems, whether they are focused on manufacturing, supply chain or energy production,
• Finance to optimize automated trading or market risk management,
• Text summarization to estimate the overall quality of a summary by extracting from a word by word logic,
• Game and recommendation engines for developing strategies in uncertain environments,
• The autonomous car to improve the vehicle’s ability to react to a given traffic event,
• …