La IA puede saber tu sexo mirando tus ojos, pero nadie sabe cómo lo hace La inteligencia artificial ha mostrado ser muy eficaz en medicina. Tanto como para detectar factores que los seres humanos somos incapaces de discernir. Por ejemplo, se ha logrado predecir el sexo de una persona, simplemente analizando la retina en una foto del fondo de su ojo. Podría parecer una banalidad si no fuese porque, a día de hoy, no se sabe cuáles son las características que diferencian la retina de hombres y mujeres. Para lograrlo se entrenó un algoritmo de aprendizaje profundo (deep learning) con 84.743 fotos del fondo de ojo procedentes del Biobanco de Reino Unido. Tanto con la validación interna como con la externa los resultados fueron muy buenos. Ojo de hombre o de mujer Fisiológicamente existen muchas diferencias entre personas con sexo biológico masculino o femenino. En el caso de los ojos también hay diferencias. Todas estas diferencias están cada vez más claras. ‘Deep learning’ para mejorar el diagnóstico de enfermedades oftalmológicas
GAN Lab: Play with Generative Adversarial Networks in Your Browser! What is a GAN? Many machine learning systems look at some kind of complicated input (say, an image) and produce a simple output (a label like, "cat"). By contrast, the goal of a generative model is something like the opposite: take a small piece of input—perhaps a few random numbers—and produce a complex output, like an image of a realistic-looking face. A generative adversarial network (GAN) is an especially effective type of generative model, introduced only a few years ago, which has been a subject of intense interest in the machine learning community. You might wonder why we want a system that produces realistic images, or plausible simulations of any other kind of data. Besides the intrinsic intellectual challenge, this turns out to be a surprisingly handy tool, with applications ranging from art to enhancing blurry images. How does a GAN work? The first idea, not new to GANs, is to use randomness as an ingredient. What's happening in the visualization? Pick a data distribution. ).
Un programme pour transformer tous les dessins des internautes en photos de chat réalistes Un site propose à ses visiteurs d’esquisser les contours du félin, puis de lui donner une allure réaliste. Un procédé qui donne parfois des résultats surprenants, voire effrayants. C’est un petit programme ingénieux qu’a fabriqué le développeur Christopher Hesse. Image-to-Image permet aux internautes de rendre leurs dessins réalistes. Bien sûr, le résultat est loin d’être parfait... mais c’est justement ce qui fait son charme, et tout son intérêt. Lire aussi : On a testé pour vous… Deep Dream, la machine à « rêves » psychédéliques de Google Deux mille photos de chat pour « entraîner » le programme Basé sur Tensorflow, une technologie d’apprentissage des machines développée par Google et accessible à tous, le programme de Christopher Hesse s’est « entraîné » sur des photos déjà existantes. Cette logique est assez proche de celle qui avait déjà amusé les internautes en 2015, quand Google avait présenté son programme Deep Dream.
DALL·E: Creating Images from Text Read paperView code DALL·E is a 12-billion parameter version of GPT-3 trained to generate images from text descriptions, using a dataset of text–image pairs. We’ve found that it has a diverse set of capabilities, including creating anthropomorphized versions of animals and objects, combining unrelated concepts in plausible ways, rendering text, and applying transformations to existing images. See also: DALL·E 2, which generates more realistic and accurate images with 4x greater resolution. an illustration of a baby daikon radish in a tutu walking a dog an armchair in the shape of an avocado. . . . a store front that has the word ‘openai’ written on it. . . . the exact same cat on the top as a sketch on the bottom GPT-3 showed that language can be used to instruct a large neural network to perform a variety of text generation tasks. Overview Like GPT-3, DALL·E is a transformer language model. Capabilities Controlling Attributes Click to edit text prompt or view more AI-generated images
Continuous video classification with TensorFlow, Inception and Recurrent Nets Part 2 of a series exploring continuous classification methods. A video is a sequence of images. In our previous post, we explored a method for continuous online video classification that treated each frame as discrete, as if its context relative to previous frames was unimportant. Today, we’re going to stop treating our video as individual photos and start treating it like the video that it is by looking at our images in a sequence. We’ll process these sequences by harnessing the magic of recurrent neural networks (RNNs). To restate the problem we outlined in our previous post: We’re attempting to continually classify video as it’s streamed, in an online system. Convolutional neural networks, which we used exclusively in our previous post, do an amazing job at taking in a fixed-size vector, like an image of an animal, and generating a fixed-size label, like the class of animal in the image. Sold! Step 2 is unique so we’ll expand on it a bit. Softmax and pool layers? Frames to sequences
Incorporated | LICEcap LICEcapsimple animated screen captures LICEcap can capture an area of your desktop and save it directly to .GIF (for viewing in web browsers, etc) or .LCF (see below). LICEcap is an intuitive but flexible application (for Windows and now OSX), that is designed to be lightweight and function with high performance. LICEcap is easy to use: view a demo (output is here). In addition to .GIF, LICEcap supports its own native lossless .LCF file format, which allows for higher compression ratios than .GIF, higher quality (more than 256 colors per frame), and more accurate timestamping. LICEcap is GPL free software, each download package includes the source. Features and options: Record directly to .GIF or .LCF. Download LICEcap v1.32 for Windows (Jun 8 2022) (250kb installer)LICEcap v1.32 for macOS (Jun 8 2022) (876kb DMG) Windows: Prevent positioning window offscreen [issue 72] Windows: sign installer/executable Source codegit clone Old versions
Recursos educativos para aplicar la Inteligencia Artificial El grupo FAIaS (Fomento de la Inteligencia Artificial en las Escuelas por sus siglas en inglés) tiene como objetivo mejorar el conocimiento de la Inteligencia Artificial (IA) en niños y jóvenes. Formado por un grupo de docentes e investigadores de diferentes países de Europa y América, el equipo FAIaS de España lo componen Gregorio Robles, Pablo Dúo, Ainhoa Erize, Luis M. Iglesias, Álvaro Molina, Cristian Ruíz, Inmaculada Caruana y Antonio J. Romero, y René Fabián Zúñiga Muñoz, docente en Colombia. Así, la IA, aunque se puede definir de diferentes formas, el grupo FAIaS considera apropiada la siguiente definición: “Capacidad de una máquina de imitar el comportamiento humano inteligente”, del Diccionario Merriam-Webster. El Machine Learning o aprendizaje automático también se puede aplicar para llevar la IA al aula. IA para océanos Quick, draw! Learning ML, Machine Learning for kids o Teachable Machine Cámara de Inteligencia Artificial Stable Diffusion, Dalle-2 y Clip Interrogator
Keras, Regression, and CNNs In this tutorial, you will learn how to train a Convolutional Neural Network (CNN) for regression prediction with Keras. You’ll then train a CNN to predict house prices from a set of images. Today is part two in our three-part series on regression prediction with Keras: Part 1: Basic regression with Keras — predicting house prices from categorical and numerical data.Part 2: Regression with Keras and CNNs — training a CNN to predict house prices from image data (today’s tutorial).Part 3: Combining categorical, numerical, and image data into a single network (next week’s tutorial). Today’s tutorial builds on last week’s basic Keras regression example, so if you haven’t read it yet make sure you go through it in order to follow along here today. By the end of this guide, you’ll not only have a strong understanding of training CNNs for regression prediction with Keras, but you’ll also have a Python code template you can follow for your own projects. Keras, Regression, and CNNs Project structure
Utiliser des images sur Internet : quelles sont les règles à respecter ? Vous cherchez des images pour illustrer vos contenus ou vos publicités ? La vigilance est de mise : des règles strictes s’appliquent. Elles dépendent du type d’image, de son utilisation, et des contenus visibles sur l’image. Pour les vidéos, les principes sont similaires. Deux types d’image : créative ou éditoriale La première règle à respecter est de sourcer l’image. Il convient ensuite de déterminer le type de l’image. Il est important de déterminer le type d’image car les règles relatives aux droits diffèrent. Deux types d’usage : éditorial ou commercial (promotionnel) S’il convient de dissocier les images éditoriales et créatives, il est également nécessaire de dissocier deux usages : l’usage éditorial et l’usage commercial. Ainsi, “les personnes prises en photographie dans le domaine public ne peuvent s’opposer à une publication dans la presse”. En revanche, si l’usage est commercial, cette autorisation est nécessaire – tout comme celle des marques et des lieux reconnaissables.