Sam 𝕏u
@SamXu03799145
Followers
188
Following
934
Media
4
Statuses
25
Research Scientist, Autodesk AI Lab, prev SFU
Canada
Joined July 2022
We are seeking a passionate and talented student with experience in 3D generation, Transformer, and Diffusion Models. In particular, if you are interested in CAD generation such as our prev work BrepGen → , please consider applying!.
0
0
1
📣 Intern Opportunity at Autodesk AI Lab! Summer 2025 | Toronto. Apply here→
2
8
70
RT @autodesk: Autodesk Research Project Bernini is new experimental generative AI that can quickly generate multiple functional 3D shapes f….
0
227
0
RT @_akhaliq: Make-A-Shape. a Ten-Million-scale 3D Shape Model. paper page: shape model, known as Make-A-Shape, ha….
0
71
0
RT @_akhaliq: Sketch-A-Shape: Zero-Shot Sketch-to-3D Shape Generation. paper page: Significant progress has recent….
0
85
0
RT @chinyich: Is there a high-level yet precise way for designers to control generative models? Can we generate layouts by simply sketching….
0
6
0
In particular, we demonstrate that our learned codebooks indeed capture the various CAD features from different hierarchy.
1
0
1
In our latest #ICML2023 paper "Hierarchical Neural Coding for Controllable CAD Model Generation" we learn the design intent from different CAD hierarchies in a self-supervised manner, and demonstrate its application on multiple tasks e.g. autocomplete.
1
4
19
RT @YasutakaFuruka1: MVDiffusion: How to take a pre-trained text2image model for a perspective view (e.g., Stable Diffusion) and retrain to….
0
37
0
RT @YasutakaFuruka1: State-of-the-art Floorplan Generation directly in vector space via continuous and discrete denoising. This afternoon p….
0
4
0
RT @ducha_aiki: “Diffusion models are general purpose learning machineries”, not only generation. Talk by . @YasutakaFuruka1 .#CVPR2023.1/….
0
41
0
RT @YizhiWang15: (1/4) Pleased to announce our.#CVPR2023 paper ARO-Net, a novel shape encoding for learning implicit field representation o….
0
7
0
RT @YasutakaFuruka1: Diffusion Models are general-purpose learning machinery. We have explored the use for reconstruction and alignment est….
0
84
0
RT @_akhaliq: Understanding Pure CLIP Guidance for Voxel Grid NeRF Models.abs: .project page: .
0
75
0
Another follow up work of dream fields () Previously I did not believe image trained models can learn good 3d. Hmm also diffusion seems to be much better than clip or is it some Google magic🤔.
openreview.net
We explore various mechanics that prevent adversarial generations from using CLIP as guidance for training a voxel grid NeRF model without any datasets.
0
0
1
RT @KarlDD: SkexGen is a new #ICML2022 paper led by @SamXu03799145 with the Autodesk AI Lab and @YasutakaFuruka1 at SFU. SkexGen (SKetch a….
0
9
0
RT @YasutakaFuruka1: #ICML2022 My first ML paper finally, presenting this Thursday. State-of-the-art CAD model generation as a construction….
0
26
0