We present CARTO, a novel approach for reconstructing multiple articulated objects from a single stereo RGB observation. We use implicit object-centric representations and learn a single geometry and articulation decoder for multiple object categories. Despite training on multiple categories, our decoder achieves a comparable reconstruction accuracy to methods that train bespoke decoders separately for each category. Combined with our stereo image encoder we infer the 3D shape, 6D pose, size, joint type, and the joint state of multiple unknown objects in a single forward pass. Our method achieves a 20.4% absolute improvement in mAP 3D IOU50 for novel instances when compared to a two-stage pipeline. Inference time is fast and can run on a NVIDIA TITAN XP GPU at 1 HZ for eight or less objects present. While only trained on simulated data, CARTO transfers to real-world object instances. Code and evaluation data is available at: http://carto.cs.uni-freiburg.de

Indie developers deserve their flowers. Not only are they responsible for amplifying fiction (and nonfiction) with immeasurable swells of immersion and creativity, but they are also proficient at submerging us in tiny pockets of humor, curiosity, and affection when we find ourselves at our worst. This year was no different. In the age of social distancing, Hades found love in a hopeless place, Among Us sussed out social anxieties via long tasks and emergency meetings, and Fall Guys filled the Mario Party–size void in our hearts by introducing humanoid jellybeans to the concept of "drip" and yeeting them through randomized elements of Takeshi's Castle. And that's just a brief synopsis of what went down in 2020. A multitude of directors, producers, animators, level designers, composers, and the like remodeled the limitations of the medium to introduce us to worlds and protagonists we never dreamed possible.

Do you need a human to create a beautiful scent? That's the question being asked as artificial intelligence (AI) starts to infiltrate the perfume industry. Companies are increasingly turning to technology in order to create more bestselling, unique fragrances that can be produced in just minutes. Last year, Swiss-based fragrance developer Givaudan Fragrances launched Carto, an artificial Intelligence-powered tool to help perfumers. Through machine learning (a way computers improve outcomes automatically by learning from past results) Carto can suggest combinations of ingredients.

In the last year, the scent industry has undergone a quiet but significant transformation thanks to artificial intelligence: In August 2018, fragrance house Firmenich announced a partnership with Swiss university Ecole Polytechnique Fédérale de Lausanne to create a digital lab in order to study AI for fragrance product development. In October 2018, another major fragrance house, Symrise, worked with IBM to develop an AI and machine learning perfumer machine called Phylra. This past April, fragrance company Givaudan launched AI platform Carto to assist perfumers in scent creation. And finally, in July, fragrance subscription startup Scentbird launched a sub-brand called Confessions of a Rebel that used AI and customer data and reviews in the creation of its four debut fragrances. Historically, perfume creation has toed the line between art and science.