The untold history of AI for the masses: Part 1: The 2010s

In the transformative decade of the 2010s, AI ceased to be an abstract concept confined to academia and industry. It became an integral part of our daily lives. From 'Content-Aware Fill' in Photoshop to virtual assistants in every phone, AI evolved from a tool into a ubiquitous companion.

Stay tuned for Part 2: The 2020s (coming very soon).

2010: Decade’s start

[30 Apr 2010] Content-Aware Fill in Photoshop CS5 (v12.0)

Adobe Photoshop CS5 revolutionizes photo editing with the introduction of Content-Aware Fill. This feature allows users to seamlessly remove and replace parts of an image, using sophisticated algorithms to fill in the space with matching background detail. It’s a game-changer for photographers and designers, simplifying complex editing tasks.

[3 May 2010] ImageNet Large Scale Visual Recognition Challenge (ILSVRC)

The ImageNet Large Scale Visual Recognition Challenge (ILSVRC) setting a benchmark in object category classification and detection. It propels the development of machine learning models that surpass human accuracy in visual tasks, leading to practical applications like autonomous vehicles and improved medical diagnostics.

[4 Nov 2010] Kinect

Microsoft’s Xbox Kinect uses computer vision to track players' movements for gaming without a controller. It combines depth sensing and machine learning to create 3D images and recognize human gestures. Its technology is based on work from the Israeli company PrimeSense. It's an example of early use of computer vision in consumer products, which has since evolved into more advanced applications.

2011

[16 Feb 2011] Watson becomes a Jeopardy! champ

IBM Watson becomes a Jeopardy! champion, marking a significant milestone in AI. Watson’s victory over human contestants showcases the system’s advanced natural language processing capabilities, with its ability to understand and respond to complex queries in a human-like manner.

[1 Sep 2011] Netflix’s recommendation system

This system analyzes user behavior and preferences to suggest personalized movie and TV show recommendations.

[4 Oct 2011] Siri

Apple unveils Siri and changing the way users interact with their devices. Siri’s integration into the iPhone 4S as a personal assistant using voice commands is essential for the normalization of AI assistants, leading to advancements in user accessibility and the development of similar technologies.

[21 Oct 2011] Clever opponents in Batman: Arkham City

Batman: Arkham City features an AI that brings a new level of challenge and realism to enemy characters, creating a more immersive and dynamic experiences for players.

2012

[Feb 2012] Deep Neural Networks (DNNs)

Deep Neural Networks (DNNs) emerge as a powerful technique for image classification. These multi-layered neural networks, inspired by the human brain, learn intricate features from raw pixel data. DNNs achieve remarkable accuracy in tasks like object recognition, paving the way for breakthroughs in computer vision and AI applications.

[7 May 2012] Content-Aware Move in Photoshop CS6 (v13.0)

Adobe unveils the impressive Content-Aware Move feature, which allows users to seamlessly relocate objects within an image while intelligently filling in the background. Additionally, the Skin Tone Detection feature can enhance portraits by automatically identifying and adjusting skin tones. These advancements empower both creative professionals and non-experts, making complex edits more accessible and efficient.

[26 Jun 2012] Google researchers train a neural network to recognize cats

Google researchers train a deep neural network on millions of YouTube images, and it spontaneously learns to recognize cats. This quirky yet significant achievement demonstrates the power of unsupervised learning and neural networks.

[9 Jul 2012] Google Now

Google launches Google Now, a predictive assistant that uses a variety of data sources to anticipate user needs. It represents a significant advancement in AI, as it proactively delivers information to users based on their habits and preferences.

[30 Sep 2012] Convolutional Neural Network (CNN) + AlexNet

AlexNet, a convolutional neural network (CNN) architecture, competes in the ImageNet competition, and achieves a top-5 error of 15.3%. This triumph in showcases the power of deep learning, particularly convolutional neural networks (CNNs), in significantly reducing error rates for image classification tasks. Its innovative architecture, which utilizes parallel GPU processing, not only outperforms existing models by a large margin but also inspires a wave of research and development in AI, leading to rapid advancements in computer vision technologies.

2013

[26 Mar 2013] Elizabeth in Bioshock Infinite

Bioshock Infinite is released, featuring Elizabeth, an AI companion that has complex emotional system. Her ability to interact dynamically with the player and the environment in a believable manner sets a new standard for non-player character characters in videogames.

[3 Apr 2013] R 3.0.0

R 3.0.0 is released, marking a major milestone for the statistical computing language. The update introduces big vectors, allowing R to better utilize memory on 64-bit systems, and represents a significant step in R’s development. It is worthy for addressing big data limitations and enhancing the capabilities of statistical analysis and visualization.

[14 Jun 2013] Clever opponents in The Last of Us

The Last of Us is released, featuring advanced enemy AI that adapts to player behavior, dynamically adjusting tactics and responses. Enemies collaborate, flank, and react realistically. This significantly enhances immersion, probably even more than the game's eye-catching visuals.

[17 June 2013] Camera Shake Reduction and Intelligent Upsampling in Photoshop CC (v14.0)

Adobe releases Photoshop CC (v14.0), introducing two significant features: Camera Shake Reduction and Intelligent Upsampling. Camera Shake Reduction minimizes blurriness caused by camera movement, improving photo quality. Intelligent Upsampling extends resolution while preserving details. These advancements truly empower photographers.

[23 Aug 2013] Clever opponents in Splinter Cell: Blacklist

Splinter Cell Blacklist launches, showcasing highly rated AI opponents. The game’s enemies adapt dynamically to the player’s actions, using cover intelligently, coordinating patrols, and responding realistically to disturbances. Their behavior challenges players strategically, enhancing the stealth genre.

[22 Nov 2013] Never Ending Image Learner (NEIL) starts learning from the internet

A computer program called the Never Ending Image Learner (NEIL) is running 24 hours a day at Carnegie Mellon University, searching the web for images, doing its best to understand them on its own and, as it builds a growing visual database, gathering common sense on a massive scale. Its ability to identify and label visuals with minimum supervision, represents a significant step forward in machine learning and computer vision.

[17 Dec 2013] Variational Autoencoders (VAEs)

Diederik Kingma and Max Welling introduce Variational Autoencoders (VAEs), a breakthrough in the field of generative models, providing a principled framework for learning deep latent-variable models and corresponding inference models.

[18 Dec 2013] Her (film)

The sci-fi movie Her hits the big screen. A sensitive and soulful man earns a living by writing personal letters for other people. Left heartbroken after his marriage ends, Theodore (Joaquin Phoenix) becomes fascinated with a new operating system which reportedly develops into an intuitive and unique entity in its own right. He starts the program and meets "Samantha" (Scarlett Johansson), whose bright voice reveals a sensitive, playful personality. Though "friends" initially, the relationship soon deepens into love.

2014

[18 Mar 2014] DeepFace

Facebook introduces DeepFace, a deep learning facial recognition system. DeepFace achieves impressive accuracy by mapping facial features to a high-dimensional space and comparing them using neural networks. It can identify faces in photos with remarkable precision, even across different lighting conditions and angles. While its capabilities raise privacy concerns, it represents a significant leap in AI-driven image analysis.

[2 Apr 2014] Cortana

Microsoft releases Cortana, their answer to Siri on iOS. Cortana combines natural language understanding, context awareness, and personalized recommendations. It assists users with tasks, reminders, and searches. While not without flaws, Cortana’s integration into Microsoft products brings virtual assistants to mainstream users.

[1 May 2014] Common Objects in Context (COCO)

Microsoft introduces the Common Objects in Context (COCO) dataset. COCO contains over 300,000 images with detailed annotations, covering 80 object categories. Researchers use it to train and evaluate object detection, segmentation, and captioning models. Its comprehensiveness and diversity advances computer vision algorithms.

[7 Jun 2014] Eugene Goostman passes the Turing test

Eugene Goostman, an AI chatbot, achieves a significant milestone by passing the Turing test during the Loebner Prize competition. The Turing test evaluates whether a machine can exhibit human-like conversation to the extent that it becomes indistinguishable from a human respondent. While some critics argue that the test conditions were not stringent enough, this achievement sparks discussions about AI’s conversational capabilities.

[10 Jun 2014] General Adversarial Network (GAN)

Ian Goodfellow introduces the Generative Adversarial Network (GAN). GANs consist of two neural networks—the generator and the discriminator—engaged in a game-like process. The generator creates realistic data (e.g., images), while the discriminator tries to distinguish between real and generated samples. GANs revolutionize generative modeling, enabling the creation of high-quality synthetic content.

[17 Sep 2014] GoogLeNet wins the ImageNet competition

The GoogLeNet architecture, developed by researchers at Google, wins the ImageNet Large Scale Visual Recognition Challenge (ILSVRC). GoogLeNet introduces the concept of inception modules, which use multiple filter sizes in parallel to capture features at different scales. This efficient and deep architecture significantly reduces computation while maintaining accuracy. GoogLeNet’s influence extends beyond image classification, inspiring subsequent neural network designs.

[3 Oct 2014] The Nemesis System in Middle-earth: Shadow of Mordor

The Nemesis System introduced in the game Middle-earth: Shadow of Mordor revolutionized game AI. It dynamically generates unique enemies with their own personalities, strengths, and weaknesses. These orcs remember player interactions, evolve, and adapt based on the player’s actions. While not directly user-facing, it showcases AI’s impact on gaming and interactive narratives.

[6 Oct 2014] Clever xenomorph in Alien: Isolation

The xenomorph aka The Alien hunts the player using sight, sound, and smell. It evokes strong tension and fear, and sets a new standard in the survival-horror genre.

[6 Nov 2014] Alexa

Amazon Alexa’s launch marks a significant milestone in voice-based AI. It brings natural language understanding and voice-controlled interactions to everyday users. Alexa’s impact extends beyond smart homes, influencing chatbots, virtual assistants, and voice interfaces across industries.

2015

[10 Apr 2015] Ex Machina (film)

The sci-fi movie Ex Machina hits the big screen. Caleb Smith (Domhnall Gleeson) a programmer at a huge Internet company, wins a contest that enables him to spend a week at the private estate of Nathan Bateman (Oscar Isaac), his firm's brilliant CEO. When he arrives, Caleb learns that he has been chosen to be the human component in a Turing test to determine the capabilities and consciousness of Ava (Alicia Vikander), a beautiful robot. However, it soon becomes evident that Ava is far more self-aware and deceptive than either man imagined.

[24 Jun 2015] OpenCV 3.0

OpenCV 3.0 is released, marking a significant advancement in computer vision capabilities. It introduces transparent API, which allows seamless offloading of computer vision routines to emerging heterogeneous computing platforms.

[1 Jul 2015] DeepDream

Google’s DeepDream becomes public, giving users a glimpse into the visual interpretations of neural networks. It kickstarts the AI art movement by transforming ordinary images into psychedelic and surreal art.

[1 Jul 2015] Spotify’s Discover Weekly

This feature utilizes AI to curate a personalized playlist of new music recommendations for each user.

[1 Sep 2015] Clever opponents in Metal Gear Solid V

Metal Gear Solid V sets a new standard for stealth gameplay with its adaptive AI system. Enemies learn from player behavior, adjusting their strategies and equipment, which compels players to continuously evolve their tactics. This advancement not only enriches the gaming experience but also showcases the potential of AI in dynamic difficulty adjustment.

[9 Nov 2015] TensorFlow

TensorFlow releases, empowering developers with a versatile framework for deep learning. It simplifies the creation of complex neural networks, making it a cornerstone in advancing AI accessibility and innovation.

2016

[5 Feb 2016] Clever opponents in XCOM 2

XCOM 2’s advanced alien tactics elevate the strategic depth of the game. The AI’s ability to challenge players with complex, unpredictable combat scenarios reflects a significant leap in game design, emphasizing the importance of tactical planning and adaptability in strategy games.

[1 Mar 2016] AlphaGo vs. world champion Go player

AlphaGo defeats the world champion Go player, showcasing the unprecedented capabilities of AI in mastering complex tasks.

[1 Mar 2016] Netflix’s intelligent personalization

Netflix announces that its recommendation system had become highly personalized, with separate algorithms tailored for each user.

[18 May 2016] Google Assistant

Google Assistant debuts, transforming voice interaction with AI. It offers a conversational experience across devices, making it easier for users to perform tasks using natural language.

[19 May 2016] Tensor Processing Units (TPUs)

Google introduces Tensor Processing Units (TPUs), specialized circuits designed to rapidly accelerate machine learning workloads, significantly impacting AI development and deployment efficiency.

[27 May 2016] GeForce 10 series

NVIDIA launches the GeForce 10 series, leveraging the Pascal architecture to enhance AI-driven graphics, setting new standards for performance and energy efficiency in gaming and AI applications.

[1 Jun 2016] StyleTransfer

Style transfer techniques emerge, allowing users to transform ordinary images into artistic masterpieces by combining the content of one image with the style of another. These algorithms enable creative expression and democratize art generation.

[20 June 2016] Content-Aware Crop in Photoshop CC 2015.5 (v17.0)

Adobe Photoshop CC 2015.5 (v17.0) introduces the Content-Aware Crop tool, which intelligently fills in missing areas when cropping images. This user-facing advancement simplifies photo editing and enhances visual compositions.

[12 Aug 2016] Complex procedural generation in No Man’s Sky

No Man’s Sky leverages complex procedural generation to create an entire universe of planets, flora, fauna, and landscapes. It demonstrates how AI-driven procedural techniques enrich game design and exploration.

[8 Sep 2016] WaveNet

WaveNet is introduced, a deep neural network capable of generating raw audio waveforms that significantly outperform traditional text-to-speech systems. It represents a leap forward in the quality of synthetic voices and music, impacting various applications from virtual assistants to music generation.

[31 Oct 2016] Clever opponents in Hitman

The AI in Hitman supports a complex simulation environment with responsive NPCs and intricate systems that enable players to become the ultimate sandbox assassin, showcasing the depth and potential of AI in creating dynamic game worlds.

[4 Nov 2016] Google Home

Google Home is released, bringing voice-activated AI into homes. It allows users to interact with services through Google Assistant, making it easier to control smart devices, access information, and manage tasks using natural language.

[11 Nov 2016] Clever opponents in Dishonored 2

Dishonored 2 introduces enemies with complex behaviors, such as the Clockwork Soldiers, which can adapt and react to the player’s actions in a variety of ways, enhancing the stealth gameplay experience.

[21 Nov 2016] Pix2Pix

Pix2Pix, an image-to-image translation framework using conditional adversarial networks, is released. It allows for a range of applications from photo editing to art creation, making AI accessible for creative tasks.

[1 Dec 2016] Zo

Microsoft's launch of the Zo chatbot marks a significant milestone in the advancement of conversational AI and natural language processing. Zo demonstrates AI's growing ability to engage in open-ended, contextually appropriate dialogue, showcasing progress in understanding and generating human-like language and personality. The chatbot's capacity to learn and evolve through interactions, as well as build personalized rapport with users, highlights the potential of AI assistants to form intelligent, socially nuanced relationships. Zo represents Microsoft's continued investment in pushing the boundaries of AI research and development, foreshadowing the future sophistication of virtual assistants and their ability to engage in intelligent conversation.

[4 Dec 2016] Westworld (TV series)

HBO’s sci-fi western series from Jonathan Nolan and Lisa Joy is inspired by the 1973 Michael Crichton film of the same name, and is set at a Wild West theme park created by Dr. Robert Ford (Anthony Hopkins) with human-like androids, where guests are encouraged to indulge their desires.

[10 Dec 2016] StackGAN

StackGAN is introduced, enabling the synthesis of photo-realistic images from text descriptions. This advancement in generative adversarial networks opens new possibilities for AI in creative design and visualization.

2017

[30 Mar 2017] CycleGAN

CycleGAN is introduced, enabling image-to-image translation without paired examples. It allows for applications like style transfer, season transfer, and photo enhancement, significantly advancing the field of unsupervised machine learning.

[1 Apr 2017] OpenPose

OpenPose releases, offering the first open-source system for real-time multi-person 2D pose detection. It enables applications in various fields such as fitness, gaming, and surveillance, making pose estimation accessible to developers.

[1 May 2017] Hulu’s sophisticated recommendation engine

This engine analyzes user data and content metadata to provide personalized show and movie suggestions.

[5 May 2017] Keras 2

Keras 2 is released, enhancing the dev-friendly deep learning library with a refined API and improved functionality. It continues to democratize AI by making deep learning more accessible and easier to implement for developers.

[17 May 2017] The Google AI division

Google announces the creation of the Google AI division, a significant move that consolidates its various AI research efforts under one umbrella. This division focuses on advancing artificial intelligence, deep learning, and developing tools to make AI work possible across various applications, including healthcare.

[1 Jun 2017] Alice & Bob

Facebook’s AI research lab trains two chatbots, Alice and Bob, to communicate with each other. They start to develop their own shorthand language, demonstrating the potential for AI systems to create more efficient means of communication, which is a noteworthy step in natural language processing research

[12 Jun 2017] Transformers

Google researchers introduce the concept of ‘Transformers’ in their influential paper “Attention Is All You Need.” This new model architecture, based solely on attention mechanisms, dispenses with recurrence and convolutions entirely, setting a new standard for natural language understanding and machine translation tasks.

[19 Jul 2017] Diffusion models

Stanford researchers introduce a groundbreaking approach to deep unsupervised learning with diffusion models. Their work, “Deep Unsupervised Learning Using Nonequilibrium Thermodynamics,” presents a novel method for data structure restoration, contributing to the development of more flexible and efficient generative models in AI.

[14 Sep 2017] Clever opponents in Divinity: Original Sin II

Gamers and critics alike have praised the AI in “Divinity: Original Sin II” for contributing to a challenging and engaging gameplay experience. It is seen as a strong point of the game, enhancing the strategic depth and providing a satisfying challenge for players.

[1 Oct 2017] Google Colab

Google Colab's release in late 2017 is a milestone because it democratizes access to powerful AI computing resources, fosters collaboration, integrates with Google's ecosystem, lowers the barrier to entry, and empowering a diverse range of devs, researchers, and students.

[4 Oct 2017] Google Lens

Google officially launches Google Lens, a powerful image recognition technology that brings up relevant information related to objects it identifies using visual analysis based on a neural network4. This tool makes AI accessible by allowing users to interact with their environment in innovative ways, such as identifying products, translating text, and understanding the world visually.

[6 Oct 2017] Blade Runner 2049 (film)

The sci-fi movie Blade Runner 2049 hits the big screen. Officer K (Ryan Gosling), a new blade runner for the Los Angeles Police Department, unearths a long-buried secret that has the potential to plunge what's left of society into chaos. His discovery leads him on a quest to find Rick Deckard (Harrison Ford), a former blade runner who's been missing for 30 years.

[18 Oct 2017] Preserve Details 2.0 in Photoshop CC 2018 (v19.0)

Adobe Photoshop CC 2018 introduces AI-powered upscaling, allowing users to enhance low-resolution images while preserving details. The “Preserve Details 2.0” algorithm leverages machine learning to intelligently resize images, making it easier for non-experts to improve image quality.

[28 Nov 2017] AttnGAN

AttnGAN, short for Attention Generative Adversarial Network, is proposed. It generates high-quality images from textual descriptions by attending to relevant parts of the input text. This advancement bridges the gap between natural language and visual content, enabling creative applications.

[31 Dec 2017] Deepfakes

A Reddit user coins the term “deepfakes” when he/she begins posting hyper-realistic AI-manipulated videos. Deepfakes use deep learning techniques, hence its name, particularly generative adversarial networks (GANs), to manipulate and create believable content.

2018

[16 Jan 2018] Alibaba’s AI vs. intellectuals

Alibaba’s language processing AI outperforms human intellect on a Stanford reading and comprehension test. This marks the first time an AI has beaten humans in such a test, signaling a significant advancement in natural language processing and machine learning.

[20 Apr 2018] Atreus in God of War

The release of “God of War” introduces Atreus, an AI companion that significantly impacts gameplay. His AI is integral to both the narrative and the player’s strategic choices, representing a leap forward in AI companionship in gaming.

[11 Jun 2018] GPT-1

GPT-1 (Generative Pre-trained Transformer 1) is released, marking a significant milestone in natural language processing. It demonstrates the power of large-scale neural networks for text generation, paving the way for subsequent models like GPT-2 and GPT-3.

[27 Jun 2018] Python 3.7

Python 3.7 introduces several improvements, including data classes, postponed evaluation of type annotations, and built-in breakpoint(). While not directly AI-related, Python remains a popular language for AI development. These enhancements benefit AI practitioners by improving code readability, maintainability, and debugging.

[9 Aug 2018] Julia 1.0

Julia 1.0 is officially released, solidifying Julia’s status as a powerful language for scientific computing, numerical analysis, and high-performance computing. Julia’s speed and ease of use make it an attractive choice for AI researchers and developers working on computationally intensive tasks such as machine learning and simulations.

[9 Aug 2018] Bixby

Samsung introduces Bixby, a virtual assistant that integrates with Samsung devices, offering voice commands, visual recognition, and language translation. It represents Samsung’s entry into the AI assistant market, competing with Siri, Google Assistant, and Alexa.

[1 Sep 2018] Amazon Prime Video’s personalization system

This system recommends content based on viewing history, ratings, and engagement across Amazon’s services.

[20 Sep 2018] GeForce 20 series

NVIDIA launches the GeForce 20 series, the first to include RTX technology, bringing real-time ray tracing and AI-driven graphics enhancements to consumer graphics cards. This series sets a new standard for high-fidelity gaming and AI applications. Notably, many artists and creators leverage these GPUs locally for generating AI art, benefiting from their powerful computational capabilities.

[28 Sep 2018] BigGAN

BigGAN is introduced, pushing the boundaries of generative adversarial networks by producing high-resolution, high-fidelity images. It achieves state-of-the-art performance on ImageNet, showcasing the potential of GANs in creative and practical AI applications.

[26 Oct 2018] Clever opponents in Red Dead Redemption 2

Red Dead Redemption 2 features an AI system that allows non-player characters (NPCs) to react and adapt to the player’s actions in a realistic manner. This level of AI sophistication contributes to an immersive and dynamic game world, setting a new benchmark for AI in open-world games.

[2 Nov 2018] BERT

Google open-sources a new technique for NLP pre-training called Bidirectional Encoder Representations from Transformers, or BERT. With this release, anyone in the world can train their own state-of-the-art question answering system (or a variety of other models) in about 30 minutes on a single Cloud TPU, or in a few hours using a single GPU.

[1 Dec 2018] Facebook’s AI versus NSFW content

Facebook starts using computer vision AI to proactively detect and filter out explicit visual content. This AI technology aims to identify objectionable content in categories such as nudity, graphic violence, and hate speech, improving the platform’s content moderation capabilities

[7 Dec 2018] PyTorch 1.0

PyTorch 1.0 is released, offering a seamless transition from research prototyping to production deployment. It includes a hybrid front end for transitioning between eager and graph execution modes, making it a valuable tool for AI developers.

[12 Dec 2018] StyleGAN

StyleGAN is introduced by Nvidia researchers, presenting a novel generative adversarial network that can produce high-quality images. It allows for intuitive control over the synthesis process and has been influential in the field of AI-generated art.

[14 Dec 2018] First game to fully support NVIDIA DLSS 1.0

Final Fantasy XV becomes the first video game to fully support NVIDIA DLSS 1.0, offering improved performance and image quality. This integration showcases the practical application of AI in enhancing gaming experiences.

[25 Jan 2019] Mr. X in Resident Evil 2

The release of Mr. X in the Resident Evil 2 remake on showcases a formidable AI antagonist. This milestone is worthy because Mr. X’s AI-driven relentless pursuit and adaptive tactics have set a new standard for horror gaming experiences, enhancing the genre’s immersion and challenge.

2019: Decade’s end

[9 Feb 2019] This Person Does Not Exist

The website ‘This Person Does Not Exist’ is created by an Uber engineer named Phillip Wang, utilizing StyleGAN software developed by Nvidia researchers to generate realistic human faces. The AI model is capable of producing a new face with each page reload, showcasing the capabilities of generative adversarial networks in creating lifelike images.

[12 Feb 2019] Enhance Details in Photoshop CC 2019 (v20.0)

Adobe Photoshop CC 2019 introduces the ‘Enhance Details’ feature, which uses machine learning to improve the quality of zoomed-in images. This AI-powered tool demonstrates practical applications of AI in professional photo editing

[14 Feb 2019] GPT-2

OpenAI releases GPT-2, a large-scale language model that significantly advances natural language processing. GPT-2 demonstrates an ability to generate coherent and contextually relevant text, impacting fields like content creation and language translation.

[15 Mar 2019] This X Does Not Exist

The website ‘This X Does Not Exist’ uses generative adversarial networks (GANs) to create realistic-looking fake versions of various objects or concepts

[18 Mar 2019] GauGAN

GauGAN, developed by NVIDIA Research, leverages generative adversarial networks (GANs) to convert rough sketches into highly realistic images. Users can draw segmentation maps and manipulate scenes, labeling segments like sand, sky, or snow. The AI model fills in details, textures, reflections, shadows, and colors based on real-world image learning. It’s like a “smart paintbrush” for creating photorealistic landscapes.

[15 Apr 2019] OpenAI Five vs. world champions in Dota 2

OpenAI Five becomes the first AI system to defeat the world champions in an esports game. It wins two back-to-back games against the Dota 2 world champion team, OG. By leveraging self-play reinforcement learning, OpenAI Five achieves superhuman performance on a challenging task, demonstrating the power of AI in complex games like Dota 2.

[16 Jul 2019] Hugging Face Transformers 1.0

Hugging Face releases Transformers 1.0, providing a comprehensive library of pre-trained models for natural language processing tasks. It simplifies the process of utilizing state-of-the-art models like BERT and GPT for developers, significantly contributing to the democratization of NLP technology.

[1 Oct 2019] Content-Aware Fill 2.0 in Photoshop CC 2020 (v21.0)

Adobe enhances Photoshop CC 2020 with an improved Content-Aware Fill feature. It allows users to remove unwanted elements from images more effectively, leveraging AI to fill in the gaps with relevant content, thus streamlining the editing process for creatives.

[1 Oct 2019] Detectron2

Facebook AI Research launches Detectron2, a PyTorch-based software system that implements state-of-the-art object detection algorithms. It is designed to support a wide range of computer vision tasks and is used in both research and production at Facebook, marking a significant advancement in the field

[25 Oct 2019] BERT + Google Search

Google implements BERT (Bidirectional Encoder Representations from Transformers) in its search algorithm, significantly improving its understanding of the nuances and context of user queries. This update affects roughly 10% of searches, marking one of the most substantial improvements to Google Search in years.

[30 Oct 2019] AlphaStar

DeepMind’s AlphaStar reaches Grandmaster level in StarCraft II, outperforming 99.8% of human players. It becomes the first AI to achieve top-tier status in an esport, using the same conditions as human players and demonstrating the potential of AI in complex strategic environments.

[4 Nov 2019] Intelligent object selection and one-click background removal in Photoshop CC 2020 (v21.0)

Adobe Photoshop CC 2020 (v21.0) introduces sophisticated features like intelligent object selection and one-click background removal, streamlining complex editing tasks.

[12 Nov 2019] Disney+'s recommendation engine

This engine analyzes user preferences and behavior to suggest personalized content.

[19 Nov 2019] Artbreeder

Artbreeder, formerly known as GANBreeder, launches as a platform that allows users to merge images to create new artworks using generative adversarial networks (GANs). It democratizes art creation, making it accessible to users without traditional artistic skills.

[3 Dec 2019] StyleGAN2

StyleGAN2 is released, addressing and improving upon the artifacts of its predecessor. This iteration offers enhanced image quality and resolution, further pushing the boundaries of AI-generated art and imagery.

---

Stay tuned for Part 2: The 2020s (coming very soon).