The instrumental variable (IV) approach is a widely used way to estimate the causal effects of a treatment on an outcome of interest from observational data with latent confounders. A standard IV is expected to be related to the treatment variable and independent of all other variables in the system. However, it is challenging to search for a standard IV from data directly due to the strict conditions. The conditional IV (CIV) method has been proposed to allow a variable to be an instrument conditioning on a set of variables, allowing a wider choice of possible IVs and enabling broader practical applications of the IV approach. Nevertheless, there is not a data-driven method to discover a CIV and its conditioning set directly from data. To fill this gap, in this paper, we propose to learn the representations of the information of a CIV and its conditioning set from data with latent confounders for average causal effect estimation. By taking advantage of deep generative models, we develop a novel data-driven approach for simultaneously learning the representation of a CIV from measured variables and generating the representation of its conditioning set given measured variables. Extensive experiments on synthetic and real-world datasets show that our method outperforms the existing IV methods.

Check out the on-demand sessions from the Low-Code/No-Code Summit to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers. The metaverse skyrocketed into our collective awareness during the height of the pandemic, when people longed for better ways to connect with each other than video calls. Gaming's hot growth during the pandemic also pushed it forward. But the metaverse became so trendy that it now faces a backlash, and folks aren't talking about it as much. Yet technologies that will power the metaverse are speeding ahead.

When GPT-3 was first launched in 2020, users were surprised with the huge performance leap from its predecessor, GPT-2. It's been over two years since OpenAI has been discreet about GPT-4--only letting out dribs of information, remaining silent for most of the time. As people have been talking about this for months, several sources hint that it's already out. Hopefully, sometime from December to February, we might be able to see the new model. Release is planned for Dec-Feb.

In 2018 at the World Economic Forum in Davos, Google CEO Sundar Pichai had something to say: "AI is probably the most important thing humanity has ever worked on. I think of it as something more profound than electricity or fire." Pichai's comment was met with a healthy dose of skepticism. AI translation is now so advanced that it's on the brink of obviating language barriers on the internet among the most widely spoken languages. College professors are tearing their hair out because AI text generators can now write essays as well as your typical undergraduate -- making it easy to cheat in a way no plagiarism detector can catch. AI-generated artwork is even winning state fairs. A new tool called Copilot uses machine learning to predict and complete lines of computer code, bringing the possibility of an AI system that could write itself one step closer.

The concept of generative artificial intelligence (GAI) poses a groundbreaking question that has until recently not been contemplated: at what stage does the relationship between humans and machines evolve from its present-day form into one that is so fundamentally changed that we can no longer regard one as being superior to the other when it comes to creative terms? Humanity stands on the brink of a new technological revolution. It is poised to harness the full potential of AI and machine learning, allowing us to automate many tasks and systems, revolutionise communication, and conserve time and money in our daily lives. Many are concerned that this could be the harbinger of a world full of robot overlords which would rob the human race of its free will. But what about those who will create those machines? In fact, some argue that in developing AI, we are creating a tool to enhance human cognition, giving us new means to think, invent and explore the universe rather than enslave humanity. Let's explore what generative AI is, where it currently stands, and where it could potentially take us in the next years. Generative AI is a branch of computer science that involves unsupervised and semi-supervised algorithms that enable computers to create new content using previously created content, such as text, audio, video, images, and code. It is all about creating authentic-looking artifacts that are completely original. In other words, generative AI is a subset of machine learning that focuses on creating algorithms that can generate new data. Generative models are used in many different application areas, from art and music to computer vision and robotics.

The breakthrough in Denoising Diffusion Probabilistic Models (DDPM) happened about 2 years ago. Since then, we observe dramatic improvements in generation tasks: GLIDE, DALL-E 2, Imagen, Stable Diffusion for images, Diffusion-LM in language modeling, diffusion for video sequences, and even diffusion for reinforcement learning. Diffusion might be the biggest trend in GraphML in 2022 -- particularly when applied to drug discovery, molecules and conformer generation, and quantum chemistry in general. Often, they are paired with the latest advancements in equivariant GNNs. Let's recapitulate the basics of diffusion models using the example of the Equivariant Diffusion paper by Hoogeboom et al using as few equations as possible The work introduces an equivariant diffusion model (EDM) for molecule generation that has to maintain E(3) equivariance over atom coordinates x (as to rotation, translation, reflection) and while node features h (such as atom types) remain invariant.

The weird and wonderful art created when AI and humans unite BBC Will AI kill art? Not likely, says the artist Alexander Reben, who has been working with AI for years. "I knew I had hit upon the right recipe when I got the following output by GPT-3 (which made me laugh a little too hard alone in my studio in lockdown):" "The sculpture contains a plunger, a toilet plunger, a plunger, a plunger, a plunger, and a plunger, each of which has been modified. The first plunger is simply a normal plunger, but the rest represent a series of plungers with more and more of the handle removed until just the rubber cup is left. The title of the artwork is "A Short History of Plungers and Other Things That Go Plunge in the Night" by the artists known as "The Plungers" (whose identity remains unknown). "The Plungers", were a collective of anonymous artists, founded in 1972. They were dedicated to the "conceptualization and promotion of a new art form called Plungism." Plungism was a creative interpretation of the idea of Plungerism, which was defined by The Plungers as "a state of mind wherein the mind of an artist is in a state of flux and able to be influenced by all things, even plungers." The Plungers' works were displayed in New York galleries and included such titles as "Plunger's Progress," "The Plungers," "The Plungers Strike Back," and "Big Plunger 4: The Final Plunger," all of which featured plungers, and "Plungers on Parade," which showed images of plungers in public spaces. The Plungers disappeared and left no trace of their identity."

A tense scene in the 2004 movie iRobot shows the character played by Will Smith arguing with an android about humanity's creative prowess. "Can a robot write a symphony?" he asks, rhetorically. "Can a robot turn a canvas into a beautiful masterpiece?" E-paper with 2-week archive so you won't miss out on content that matters to you Join ST's Telegram channel and get the latest breaking news delivered to you.

OpenAI wanted to advance artificial intelligence (AI) and machine learning research in a more creative way, so they trained their new bot to play Minecraft using over 70,000 hours of gameplay footage from YouTube. The bot utilized the gameplay actions and tutorials to learn how to execute complex in-game sequences that would take a normal player around 24,000 individual actions to accomplish. Their behavioral cloning model accomplished many tasks including learning how to chop down trees to collect logs and then craft those into planks. This sequence would typically take a human Minecraft player approximately 50 seconds or 1,000 consecutive game actions. Additionally, the model performs other complex skills humans often do in the game, such as swimming, hunting animals for food, and eating that food.

Structure-based drug design (SBDD) aims to discover drug candidates by finding molecules (ligands) that bind tightly to a disease-related protein (targets), which is the primary approach to computer-aided drug discovery. Recently, applying deep generative models for three-dimensional (3D) molecular design conditioned on protein pockets to solve SBDD has attracted much attention, but their formulation as probabilistic modeling often leads to unsatisfactory optimization performance. On the other hand, traditional combinatorial optimization methods such as genetic algorithms (GA) have demonstrated state-of-the-art performance in various molecular optimization tasks. However, they do not utilize protein target structure to inform design steps but rely on a random-walk-like exploration, which leads to unstable performance and no knowledge transfer between different tasks despite the similar binding physics. To achieve a more stable and efficient SBDD, we propose Reinforced Genetic Algorithm (RGA) that uses neural models to prioritize the profitable design steps and suppress random-walk behavior. The neural models take the 3D structure of the targets and ligands as inputs and are pre-trained using native complex structures to utilize the knowledge of the shared binding physics from different targets and then fine-tuned during optimization. We conduct thorough empirical studies on optimizing binding affinity to various disease targets and show that RGA outperforms the baselines in terms of docking scores and is more robust to random initializations. The ablation study also indicates that the training on different targets helps improve performance by leveraging the shared underlying physics of the binding processes. The code is available at https://github.com/futianfan/reinforced-genetic-algorithm.