My sense is that if you spend the time to deeply research a market, talk to customers and their current solutions (putative competitors) to find something that isn’t “1st order” / or demo’d on twitter (ie is esoteric but actually massive), a lot of these concerns/tips go away.
RESEARCH
-
NAVI: High-Quality 3D Shape and Pose Annotations for Images
By
–
NAVI: Category-Agnostic Image Collections with High-Quality 3D Shape and Pose Annotations paper page: https://
huggingface.co/papers/2306.09
109
… Recent advances in neural reconstruction enable high-quality 3D object reconstruction from casually captured image collections. Current techniques -
Agile Catching with Whole-Body MPC and Blackbox Policy Learning
By
–
Agile Catching with Whole-Body MPC and Blackbox Policy Learning
— AK (@_akhaliq) 16 juin 2023
paper page: https://t.co/A8ZuUrvwnP
address a benchmark task in agile robotics: catching objects thrown at high-speed. This is a challenging task that involves tracking, intercepting, and cradling a thrown object… pic.twitter.com/mKCyUxOIGLAgile Catching with Whole-Body MPC and Blackbox Policy Learning paper page: https://
huggingface.co/papers/2306.08
205
… address a benchmark task in agile robotics: catching objects thrown at high-speed. This is a challenging task that involves tracking, intercepting, and cradling a thrown object -
DreamHuman: Generate Animatable 3D Avatars from Text
By
–
DreamHuman: Animatable 3D Avatars from Text
— AK (@_akhaliq) 16 juin 2023
paper page: https://t.co/w3dnN9wS6o
present DreamHuman, a method to generate realistic animatable 3D human avatar models solely from textual descriptions. Recent text-to-3D methods have made considerable strides in generation, but are… pic.twitter.com/HpsWFJnsxrDreamHuman: Animatable 3D Avatars from Text paper page: https://
huggingface.co/papers/2306.09
329
… present DreamHuman, a method to generate realistic animatable 3D human avatar models solely from textual descriptions. Recent text-to-3D methods have made considerable strides in generation, but are -
Neural Relighting with Subsurface Scattering via Radiance Transfer
By
–
Neural Relighting with Subsurface Scattering by Learning the Radiance Transfer Gradient paper page: https://
huggingface.co/papers/2306.09
322
… Reconstructing and relighting objects and scenes under varying lighting conditions is challenging: existing neural rendering methods often cannot handle -
Humans as biological machines: trillion cellular networks
By
–
You and I both are miracles of trillions of little machines all bundled together. https://t.co/YlDb2cqfkp
— Robert Scoble (@Scobleizer) 16 juin 2023You and I both are miracles of trillions of little machines all bundled together.
-
Diffusion Models for Zero-Shot Open-Vocabulary Segmentation
By
–
Diffusion Models for Zero-Shot Open-Vocabulary Segmentation
— AK (@_akhaliq) 16 juin 2023
paper page: https://t.co/JP6BymlyTs
The variety of objects in the real world is nearly unlimited and is thus impossible to capture using models trained on a fixed set of categories. As a result, in recent years,… pic.twitter.com/BVI1gSpjRnDiffusion Models for Zero-Shot Open-Vocabulary Segmentation paper page: https://
huggingface.co/papers/2306.09
316
… The variety of objects in the real world is nearly unlimited and is thus impossible to capture using models trained on a fixed set of categories. As a result, in recent years, -
TryOnDiffusion: Virtual Garment Try-On Using Dual UNet Architecture
By
–
TryOnDiffusion: A Tale of Two UNets
— AK (@_akhaliq) 16 juin 2023
paper page: https://t.co/SuDRe7rqQD
Given two images depicting a person and a garment worn by another person, our goal is to generate a visualization of how the garment might look on the input person. A key challenge is to synthesize a… pic.twitter.com/qcIXtUbqyLTryOnDiffusion: A Tale of Two UNets paper page: https://
huggingface.co/papers/2306.08
276
… Given two images depicting a person and a garment worn by another person, our goal is to generate a visualization of how the garment might look on the input person. A key challenge is to synthesize a -
DORSal: Diffusion for Object-centric Scene Representations
By
–
DORSal: Diffusion for Object-centric Representations of Scenes et al. paper page: https://
huggingface.co/papers/2306.08
068
… Recent progress in 3D scene understanding enables scalable learning of representations across large datasets of diverse scenes. As a consequence, generalization to unseen -
Anticipatory Music Transformer: Controllable Generative Model for Temporal Point Processes
By
–
Anticipatory Music Transformer paper page: https://
huggingface.co/papers/2306.08
620
… introduce anticipation: a method for constructing a controllable generative model of a temporal point process (the event process) conditioned asynchronously on realizations of a second, correlated process (the