CS348n Course Schedule

CS348n Student Presentation Papers

Please select papers to present among those listed for topic/data show,

Reacall that these are shifted to be one lectutre after the main lecture on the same topic.

Each student should present 1-2 papers, 8 mins max (+ 2 min questions).

Presenters in the same session should negotiate with each other to avoid paper overlaps.

Monday

Wednesday

April 3	April 5

April 10	April 12
	Neural Architectures for Geometric Data: Point Clouds. [Speakers provided by the class staff.] KPConv: Flexible and Deformable Convolution for Point Clouds Point Convolutional Neural Networks by Extension Operators An End-to-End Transformer Model for 3D Object Detection (3DETR) Point-Voxel CNN for Efficient 3D Deep Learning Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework Generative PointNet: Deep Energy-Based Learning on Unordered Point Sets for 3D Generation, Reconstruction and Classification
April 17	April 19
Generative Models: VAEs and Coordinate-Based Networks. Implicit Autoencoder for Point Cloud Self-supervised Representation Learning Deep Structured Implicit Functions Local Implicit Grid Representations for 3D Scenes Neural Wavelet-domain Diffusion for 3D shape generation GenSDF: Two-stage learning of generalizable signed distance functions AutoSDF: Shape Priors for 3D Completion, Reconstruction and Generation Auto-Encoding Variational Bayes DualSDF: Semantic Shape Manipulation using a Two-Level Representation Shape As Points: A Differentiable Poisson Solver	3D Public Data Sets. Flow and Auto-Regressive Models. Normalizing Flows: An Introduction and Review of Current Methods Learning Gradient Fields for Shape Generation Common Objects in 3D: Large-Scale Learning and Evaluation of Real-life 3D Category Reconstruction Objaverse: A Universe of Annotated 3D Objects 3DCoMPaT: Composition of Materials on Parts of 3D Things ProcTHOR: Large-Scale Embodied AI Using Procedural Generation
April 24	April 26
Parametric Models. Generative Adversarial Networks (GANs). ParSeNet: A Parametric Surface Fitting Network for 3D Point Clouds Analyzing and Improving the Image Quality of StyleGAN Alias-Free Generative Adversarial Networks Efficient Geometry-aware 3D Generative Adversarial Networks (EG3D) Compositional Visual Generation with Composable Diffusion Models SP-GAN: Sphere-Guided 3D Shape Generation and Manipulation PQ-NET: A Generative Part Seq2Seq Network for 3D Shapes Mimic3D: Thriving 3D-Aware GANs via 3D-to-2D Imitation	Introduction to Denoising Diffusion Models. Classifier-free Diffusion Guidance Imagen Stable Diffusion Diffusion Models Beat GANs on Image Synthesis High-Resolution Image Synthesis with Latent Diffusion Models Structured Denoising Diffusion Models in Discrete State-Spaces LION: Latent Point Diffusion Models for 3D Shape Generation 3D Neural Field Generation using Triplane Diffusion 3D Shape Generation and Completion through Point-Voxel Diffusion 3DShape2VecSet: A 3D Shape Representation for Neural Fields and Generative Diffusion Models Diffusion-Based Point Cloud Generation With Smoothness Constraints Diffusion Probabilistic Models for 3D Point Cloud Generation
May1	May 3
Hierarchical Generation of Structure and Geometry. Composite Shape Modeling via Latent Space Factorization SDM-NET: Deep Generative Network for Structured Deformable Mesh SCORES: Shape Composition with Recursive Substructure Priors SPAGHETTI: Editing Implicit Shapes Through Part Aware Generation	Vector Graphics, Deep Architectures for Meshes. Deep Sketch-Based Modeling of Man-Made Shapes DeepCAD: A Deep Generative Network for Computer-Aided Design Models PIE-NET: Parametric Inference of Point Cloud Edges Neural Template: Topology-aware Reconstruction and Disentangled Generation of 3D Meshes HouseDiffusion: Vector Floorplan Generation via a Diffusion Model with Discrete and Continuous Denoising Controllable Mesh Generation Through Sparse Latent Point Diffusion Models MeshDiffusion: Score-based Generative 3D Mesh Modeling
May 8	May 10
Pose Equivariance and Invariance in 3D Data. ConDor: Self-Supervised Canonicalization of 3D Pose for Partial Shapes CNNs on Surfaces using Rotation-Equivariant Features Equivariant Multi-View Networks SE(3)-Equivariant Attention Networks for Shape Reconstruction in Function Space Shape-Pose Disentanglement using SE(3)-equivariant Vector Neurons Learning SO(3) Equivariant Representations with Spherical CNNs	Conditional Generation: 3D Shape Completion. PU-GAN: a Point Cloud Upsampling Adversarial Network Multimodal Shape Completion via Conditional Generative Adversarial Networks Unsupervised 3D Shape Completion through GAN Inversion ShapeFormer: Transformer-based Shape Completion via Sparse Representation Deep Marching Tetrahedra: a Hybrid Representation for High-Resolution 3D Shape Synthesis SDFusion: Multimodal 3D Shape Completion, Reconstruction, and Generation Neural Fields as Learnable Kernels for 3D Reconstruction 3D Shape Generation and Completion through Point-Voxel Diffusion
May 15	May 17
Neural Rendering. Neural Radiance Fields (NeRFs) I. GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields pixelNeRF: Neural Radiance Fields from One or Few Images nerf2nerf: Pairwise Registration of Neural Radiance Fields DiffusioNeRF- Regularizing Neural Radiance Fields with Denoising Diffusion Models RobustNeRF: Ignoring Distractors with Robust Losses Editing Conditional Radiance Fields DiffRF: Rendering-Guided 3D Radiance Field Diffusion	Conditional Generation: From Image to 3D Shape. D^2IM-Net: Learning Detail Disentangled Implicit Fields from Single Images CoReNet: Coherent 3D scene reconstruction from a single RGB image Escaping Plato's Cave: 3D Shape From Adversarial Rendering Occupancy Networks: Learning 3D Reconstruction in Function Space MagicPony: Learning Articulated 3D Animals in the Wild Hi-LASSIE: High-Fidelity Articulated Shape and Skeleton Discovery from Sparse Image Ensemble Learning Deformable Tetrahedral Meshes for 3D Reconstruction NerfDiff: Single-image View Synthesis with NeRF-guided Distillation RealFusion: 360 Reconstruction of Any Object from a Single Image HOLODIFFUSION: Training a 3D Diffusion Model using 2D Images
May 22	May 24
Object/Scene Generation and Language I. Dream3D: Zero-Shot Text-to-3D Synthesis Using 3D Shape Prior and Text-to-Image Diffusion Models ShapeCrafter: A Recursive Text-Conditioned 3D Shape Generation Model CLIP-Forge: Towards Zero-Shot Text-To-Shape Generation Zero-Shot Text-Guided Object Generation with Dream Fields Diffusion-SDF: Text-to-Shape via Voxelized Diffusion GET3D: A Generative Model of High Quality 3D Textured Shapes Learned from Images SceneScape: Text-Driven Consistent Scene Generation Point-E: A System for Generating 3D Point Clouds from Complex Prompts Shap-E: Generating Conditional 3D Implicit Functions	Object/Scene Generation and Language II. DiffusionSDF: Conditional Generative Modeling of Signed Distance Functions Diffusion Probabilistic Models for Scene-Scale 3D Categorical Data SparseFusion: Distilling View-conditioned Diffusion for 3D Reconstruction DiffRF: Rendering-Guided 3D Radiance Field Diffusion Tetrahedral Diffusion Models for 3D Shape Generation Magic3D: High-Resolution Text-to-3D Content Creation RenderDiffusion: Image Diffusion for 3D Reconstruction, Inpainting and Generation
May 29	May 31
Memorial Day (holiday, no classes).	Shape Edits/Deformations. 3DN: 3D Deformation Network CAD-DEFORM: DEFORMABLE FITTING OF CAD MODELS TO 3D SCANS LOGAN: Unpaired Shape Transform in Latent Overcomplete Space KeypointDeformer: Unsupervised 3D Keypoint Discovery for Shape Control Cadex: Learning canonical deformation coordinate space for dynamic surface representation via neural homeomorphism Neural Shape Deformation Priors NeuralMLS: Geometry-Aware Control Point Deformation GeoCode: Interpretable Shape Programs
June 5	June 7
Neural Radiance Fields (NeRFs) II. D-NeRF: Neural Radiance Fields for Dynamic Scenes Instant Neural Graphics Primitives with a Multiresolution Hash Encoding (Instant NGP) Learning Object-Compositional Neural Radiance Field for Editing Scene Rendering Editing Conditional Radiance Fields Plenoxels: Radiance Fields without Neural Networks KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs BANMo: Building Animatable 3D Neural Models from Many Casual Videos Panoptic Neural Fields: A Semantic Object-Aware Neural Scene Representation Few-View Object Reconstruction with Unknown Categories and Camera Poses CLIP-NeRF: Text-and-Image Driven Manipulation of Neural Radiance Fields