We adapt alignment techniques from reasoning LLMs to the task of generating engineering sketch constraints found in computer-aided design (CAD) models. Engineering sketches consist of geometric primitives (e.g. points, lines) connected by constraints (e.g. perpendicular, tangent) that define the relationships between them. For a design to be easily editable, the constraints must effectively capture design intent, ensuring the geometry updates predictably when parameters change. Although current approaches can generate CAD designs, an open challenge remains to align model outputs with design intent, we label this problem 'design alignment'. A critical first step towards aligning generative CAD models is to generate constraints which fully-constrain all geometric primitives, without over-constraining or distorting sketch geometry. Using alignment techniques to train an existing constraint generation model with feedback from a constraint solver, we are able to fully-constrain 93% of sketches compared to 34% when using a naive supervised fine-tuning (SFT) baseline and only 8.9% without SFT. Our approach can be applied to any existing constraint generation model and sets the stage for further research bridging alignment strategies between the language and design domains. Additional results can be found at https://autodeskailab.github.io/aligning-constraint-generation/.

Realizing a designer’s intent in software currently requires tedious manipulation of geometric primitives, such as points and curves. By contrast, designers routinely communicate more abstract design goals to one another using an efficient combination of natural language and drawings. What would it take to develop artificial systems that understand how humans naturally convey design intent, and thereby enable more seamless interactions between humans and machines throughout the design process? First, it is vital to establish benchmarks that showcase the full range of strategies that humans use to successfully communicate about design intent. Here we take initial steps towards that goal by conducting an online study in which pairs of human participants – a “Designer” and “Maker” – collaborated over multiple turns to recreate target designs. In each turn, Designers sent messages containing language, drawings, or both to the Maker, describing how to modify an existing design toward the target. We found a preference for communicating using drawings in early turns and observed several multimodal strategies for conveying design intent. By comparing how human Makers and GPT-4V carried out instructions, we identify a gap in human and machine understanding of multimodal instructions and suggest a path for bridging this gap.

The automated assembly of complex products requires a system that can automatically plan a physically feasible sequence of actions for assembling many parts together. In this paper, we present ASAP, a physics-based planning approach for automatically generating such a sequence for general-shaped assemblies. ASAP accounts for gravity to design a sequence where each sub-assembly is physically stable with a limited number of parts being held and a support surface. We apply efficient tree search algorithms to reduce the combinatorial complexity of determining such an assembly sequence. The search can be guided by either geometric heuristics or graph neural networks trained on data with simulation labels. Finally, we show the superior performance of ASAP at generating physically realistic assembly sequence plans on a large dataset of hundreds of complex product assemblies. We further demonstrate the applicability of ASAP on both simulation and real-world robotic setups. Project website: asap.csail.mit.edu

This paper presents BrepGen , a diffusion-based generative approach that directly outputs a Boundary representation (B-rep) Computer-Aided Design (CAD) model. BrepGen represents a B-rep model as a novel structured latent geometry in a hierarchical tree. With the root node representing a whole CAD solid, each element of a B-rep model (i.e., a face, an edge, or a vertex) progressively turns into a child-node from top to bottom. B-rep geometry information goes into the nodes as the global bounding box of each primitive along with a latent code describing the local geometric shape. The B-rep topology information is implicitly represented by node duplication. When two faces share an edge, the edge curve will appear twice in the tree, and a T-junction vertex with three incident edges appears six times in the tree with identical node features. Starting from the root and progressing to the leaf, BrepGen employs Transformer-based diffusion models to sequentially denoise node features while duplicated nodes are detected and merged, recovering the B-Rep topology information. Extensive experiments show that BrepGen advances the task of CAD B-rep generation, surpassing existing methods on various benchmarks. Results on our newly collected furniture dataset further showcase its exceptional capability in generating complicated geometry. While previous methods were limited to generating simple prismatic shapes, BrepGen incorporates free-form and doubly-curved surfaces for the first time. Additional applications of BrepGen include CAD autocomplete and design interpolation. The code, pretrained models, and dataset are available at https://github.com/samxuxiang/BrepGen.

Recent works have demonstrated that natural language can be used to generate and edit 3D shapes. However, these methods generate shapes with limited fidelity and diversity. We introduce CLIP-Sculptor, a method to address these constraints by producing high-fidelity and diverse 3D shapes without the need for (text, shape) pairs during training. CLIP-Sculptor achieves this in a multi-resolution approach that first generates in a low-dimensional latent space and then upscales to a higher resolution for improved shape fidelity. For improved shape diversity, we use a discrete latent space which is modeled using a transformer conditioned on CLIP's image-text embedding space. We also present a novel variant of classifier-free guidance, which improves the accuracy-diversity trade-off. Finally, we perform extensive experiments demonstrating that CLIP-Sculptor outperforms state-of-the-art baselines.

This paper presents a novel generative model for Computer Aided Design (CAD) that 1) represents high-level design concepts of a CAD model as a three-level hierarchical tree of neural codes, from global part arrangement down to local curve geometry; and 2) controls the generation or completion of CAD models by specifying the target design using a code tree. Concretely, a novel variant of a vector quantized VAE with "masked skip connection" extracts design variations as neural codebooks at three levels. Two-stage cascaded auto-regressive transformers learn to generate code trees from incomplete CAD models and then complete CAD models following the intended design. Extensive experiments demonstrate superior performance on conventional tasks such as unconditional generation while enabling novel interaction capabilities on conditional generation tasks. The code is available at https://github.com/samxuxiang/hnc-cad.

Abstract Procedural models (i.e. symbolic programs that output visual data) are a historically‐popular method for representing graphics content: vegetation, buildings, textures, etc. They offer many advantages: interpretable design parameters, stochastic variations, high‐quality outputs, compact representation, and more. But they also have some limitations, such as the difficulty of authoring a procedural model from scratch. More recently, AI‐based methods, and especially neural networks, have become popular for creating graphic content. These techniques allow users to directly specify desired properties of the artifact they want to create (via examples, constraints, or objectives), while a search, optimization, or learning algorithm takes care of the details. However, this ease of use comes at a cost, as it's often hard to interpret or manipulate these representations. In this state‐of‐the‐art report, we summarize research on neurosymbolic models in computer graphics: methods that combine the strengths of both AI and symbolic programs to represent, generate, and manipulate visual data. We survey recent work applying these techniques to represent 2D shapes, 3D shapes, and materials & textures. Along the way, we situate each prior work in a unified design space for neurosymbolic models, which helps reveal underexplored areas and opportunities for future research.

Physical products are often complex assemblies combining a multitude of 3D parts modeled in computer-aided design (CAD) software. CAD designers build up these assemblies by aligning individual parts to one another using constraints called joints. In this paper we introduce JoinABLe, a learning-based method that assembles parts together to form joints. JoinABLe uses the weak supervision available in standard parametric CAD files without the help of object class labels or human guidance. Our results show that by making network predictions over a graph representation of solid models we can outperform multiple baseline methods with an accuracy (79.53%) that approaches human performance (80%). Finally, to support future research we release the Fusion 360 Gallery assembly dataset, containing assemblies with rich information on joints, contact surfaces, holes, and the underlying assembly graph structure.

We present SkexGen, a novel autoregressive generative model for computer-aided design (CAD) construction sequences containing sketch-and-extrude modeling operations. Our model utilizes distinct Transformer architectures to encode topological, geometric, and extrusion variations of construction sequences into disentangled codebooks. Autoregressive Transformer decoders generate CAD construction sequences sharing certain properties specified by the codebook vectors. Extensive experiments demonstrate that our disentangled codebook representation generates diverse and high-quality CAD models, enhances user control, and enables efficient exploration of the design space. The code is available at https://samxuxiang.github.io/skexgen.

User modeling is crucial to understanding user behavior and essential for improving user experience and personalized recommendations. When users interact with software, vast amounts of command sequences are generated through logging and analytics systems. These command sequences contain clues to the users’ goals and intents. However, these data modalities are highly unstructured and unlabeled, making it difficult for standard predictive systems to learn from. We propose SimCURL, a simple yet effective contrastive self-supervised deep learning framework that learns user representation from unlabeled command sequences. Our method introduces a user-session network architecture, as well as session dropout as a novel way of data augmentation. We train and evaluate our method on a real-world command sequence dataset of more than half a billion commands. Our method shows significant improvement over existing methods when the learned representation is transferred to downstream tasks such as experience and expertise classification.

Reverse Engineering a CAD shape from other representations is an important geometric processing step for many downstream applications. In this work, we introduce a novel neural network architecture to solve this challenging task and approximate a smoothed signed distance function with an editable, constrained, prismatic CAD model. During training, our method reconstructs the input geometry in the voxel space by decomposing the shape into a series of 2D profile images and 1D envelope functions. These can then be recombined in a differentiable way allowing a geometric loss function to be defined. During inference, we obtain the CAD data by first searching a database of 2D constrained sketches to find curves which approximate the profile images, then extrude them and use Boolean operations to build the final CAD model. Our method approximates the target shape more closely than other methods and outputs highly editable constrained parametric sketches which are compatible with existing CAD software.

Assembly planning is the core of automating product assembly, maintenance, and recycling for modern industrial manufacturing. Despite its importance and long history of research, planning for mechanical assemblies when given the final assembled state remains a challenging problem. This is due to the complexity of dealing with arbitrary 3D shapes and the highly constrained motion required for real-world assemblies. In this work, we propose a novel method to efficiently plan physically plausible assembly motion and sequences for real-world assemblies. Our method leverages the assembly-by-disassembly principle and physics-based simulation to efficiently explore a reduced search space. To evaluate the generality of our method, we define a large-scale dataset consisting of thousands of physically valid industrial assemblies with a variety of assembly motions required. Our experiments on this new benchmark demonstrate we achieve a state-of-the-art success rate and the highest computational efficiency compared to other baseline algorithms. Our method also generalizes to rotational assemblies (e.g., screws and puzzles) and solves 80-part assemblies within several minutes.

We introduce UV-Net, a novel neural network architecture and representation designed to operate directly on Boundary representation (B-rep) data from 3D CAD models. The B-rep format is widely used in the design, simulation and manufacturing industries to enable sophisticated and precise CAD modeling operations. However, B-rep data presents some unique challenges when used with modern machine learning due to the complexity of the data structure and its support for both continuous non-Euclidean geometric entities and discrete topological entities. In this paper, we propose a unified representation for B-rep data that exploits the U and V parameter domain of curves and surfaces to model geometry, and an adjacency graph to explicitly model topology. This leads to a unique and efficient network architecture, UV-Net, that couples image and graph convolutional neural networks in a compute and memory-efficient manner To aid in future research we present a synthetic labelled B-rep dataset, SolidLetters, derived from human designed fonts with variations in both geometry and topology. Finally we demonstrate that UV-Net can generalize to supervised and unsupervised tasks on five datasets, while outperforming alternate 3D shape representations such as point clouds, voxels, and meshes.

Boundary representation (B-rep) models are the standard way 3D shapes are described in Computer-Aided Design (CAD) applications. They combine lightweight parametric curves and surfaces with topological information which connects the geometric entities to describe manifolds. In this paper we introduce BRepNet, a neural network architecture designed to operate directly on B-rep data structures, avoiding the need to approximate the model as meshes or point clouds. BRepNet defines convolutional kernels with respect to oriented coedges in the data structure. In the neighborhood of each coedge, a small collection of faces, edges and coedges can be identified and patterns in the feature vectors from these entities detected by specific learnable parameters. In addition, to encourage further deep learning research with B-reps, we publish the Fusion 360 Gallery segmentation dataset. A collection of over 35,000 B-rep models annotated with information about the modeling operations which created each face. We demonstrate that BRepNet can segment these models with higher accuracy than methods working on meshes, and point clouds.

In computer-aided design (CAD), the ability to "reverse engineer" the modeling steps used to create 3D shapes is a long-sought-after goal. This process can be decomposed into two sub-problems: converting an input mesh or point cloud into a boundary representation (or B-rep), and then inferring modeling operations which construct this B-rep. In this paper, we present a new system for solving the second sub-problem. Central to our approach is a new geometric representation: the zone graph. Zones are the set of solid regions formed by extending all B-Rep faces and partitioning space with them; a zone graph has these zones as its nodes, with edges denoting geometric adjacencies between them. Zone graphs allow us to tractably work with industry-standard CAD operations, unlike prior work using CSG with parametric primitives. We focus on CAD programs consisting of sketch + extrude + Boolean operations, which are common in CAD practice. We phrase our problem as search in the space of such extrusions permitted by the zone graph, and we train a graph neural network to score potential extrusions in order to accelerate the search. We show that our approach outperforms an existing CSG inference baseline in terms of geometric reconstruction accuracy and reconstruction time, while also creating more plausible modeling sequences.

Parametric computer-aided design (CAD) is a standard paradigm used to design manufactured objects, where a 3D shape is represented as a program supported by the CAD software. Despite the pervasiveness of parametric CAD and a growing interest from the research community, currently there does not exist a dataset of realistic CAD models in a concise programmatic form. In this paper we present the Fusion 360 Gallery , consisting of a simple language with just the sketch and extrude modeling operations, and a dataset of 8,625 human design sequences expressed in this language. We also present an interactive environment called the Fusion 360 Gym , which exposes the sequential construction of a CAD program as a Markov decision process, making it amendable to machine learning approaches. As a use case for our dataset and environment, we define the CAD reconstruction task of recovering a CAD program from a target geometry. We report results of applying state-of-the-art methods of program synthesis with neurally guided search on this task.

This work presents a method for 3D printing hair-like structures on both flat and curved surfaces. It allows a user to design and fabricate hair geometries that are smaller than 100 micron. We built a software platform to let users quickly define the hair angle, thickness, density, and height. The ability to fabricate customized hair-like structures not only expands the library of 3D-printable shapes, but also enables us to design passive actuators and swipe sensors. We also present several applications that show how the 3D-printed hair can be used for designing everyday interactive objects.

HideOut is a mobile projector-based system that enables new applications and interaction techniques with tangible objects and surfaces. HideOut uses a device mounted camera to detect hidden markers applied with infrared-absorbing ink. The obtrusive appearance of fiducial markers is avoided and the hidden marker surface doubles as a functional projection surface. We present example applications that demonstrate a wide range of interaction scenarios, including media navigation tools, interactive storytelling applications, and mobile games. We explore the design space enabled by the HideOut system and describe the hidden marker prototyping process. HideOut brings tangible objects to life for interaction with the physical world around us.

We present Lumitrack, a novel motion tracking technology that uses projected structured patterns and linear optical sensors. Each sensor unit is capable of recovering 2D location within the projection area, while multiple sensors can be combined for up to six degree of freedom (DOF) tracking. Our structured light approach is based on special patterns, called m-sequences, in which any consecutive sub-sequence of m bits is unique. Lumitrack can utilize both digital and static projectors, as well as scalable embedded sensing configurations. The resulting system enables high-speed, high precision, and low-cost motion tracking for a wide range of interactive applications. We detail the hardware, operation, and performance characteristics of our approach, as well as a series of example applications that highlight its immediate feasibility and utility.

We introduce InfraStructs , material-based tags that embed information inside digitally fabricated objects for imaging in the Terahertz region. Terahertz imaging can safely penetrate many common materials, opening up new possibilities for encoding hidden information as part of the fabrication process. We outline the design, fabrication, imaging, and data processing steps to fabricate information inside physical objects. Prototype tag designs are presented for location encoding, pose estimation, object identification, data storage, and authentication. We provide detailed analysis of the constraints and performance considerations for designing InfraStruct tags. Future application scenarios range from production line inventory, to customized game accessories, to mobile robotics.

We present an approach to 3D printing custom optical elements for interactive devices labelled Printed Optics. Printed Optics enable sensing, display, and illumination elements to be directly embedded in the casing or mechanical structure of an interactive device. Using these elements, unique display surfaces, novel illumination techniques, custom optical sensors, and embedded optoelectronic components can be digitally fabricated for rapid, high fidelity, highly customized interactive devices. Printed Optics is part of our long term vision for interactive devices that are 3D printed in their entirety. In this paper we explore the possibilities for this vision afforded by fabrication of custom optical elements using today's 3D printing technology.

Icons in graphical user interfaces convey information in a mostly universal fashion that allows users to immediately interact with new applications, systems and devices. In this paper, we define Kineticons - an iconographic scheme based on motion. By motion, we mean geometric manipulations applied to a graphical element over time (e.g., scale, rotation, deformation). In contrast to static graphical icons and icons with animated graphics, kineticons do not alter the visual content or "pixel-space" of an element. Although kineticons are not new - indeed, they are seen in several popular systems - we formalize their scope and utility. One powerful quality is their ability to be applied to GUI elements of varying size and shape from a something as small as a close button, to something as large as dialog box or even the entire desktop. This allows a suite of system-wide kinetic behaviors to be reused for a variety of uses. Part of our contribution is an initial kineticon vocabulary, which we evaluated in a 200 participant study. We conclude with discussion of our results and design recommendations.

We present the MotionBeam metaphor for character interaction with handheld projectors. Our work draws from the tradition of pre-cinema handheld projectors that use direct physical manipulation to control projected imagery. With our prototype system, users interact and control projected characters by moving and gesturing with the handheld projector itself. This creates a unified interaction style where input and output are tied together within a single device. We introduce a set of interaction principles and present prototype applications that provide clear examples of the MotionBeam metaphor in use. Finally we describe observations and insights from a preliminary user study with our system.

I present a summary of my research dealing with character interaction using handheld projectors. My work draws from the tradition of pre-cinema handheld projectors that use direct physical manipulation to control projected imagery. I build upon this work with a system allowing users to interactively control characters by moving and gesturing with the handheld projector itself. This creates a unified interaction style where input and output are tied together within a single device. I present a prototype handheld platform, several games, and augmented reality application scenarios to illustrate the approach.

We present a series of prototype devices that use real-time input to fabricate physical form: Interactive Fabrication. Our work maps out the problem space of real-time control for digital fabrication devices, and examines where alternative interfaces for digital fabrication are relevant. We conclude by reflecting upon the potential of interactive fabrication and outline a number of considerations for future research in this area.

We introduce SideBySide, a system designed for ad-hoc multi-user interaction with handheld projectors. SideBySide uses device-mounted cameras and hybrid visible/infrared light projectors to track multiple independent projected images in relation to one another. This is accomplished by projecting invisible fiducial markers in the near-infrared spectrum. Our system is completely self-contained and can be deployed as a handheld device without instrumentation of the environment. We present the design and implementation of our system including a hybrid handheld projector to project visible and infrared light, and techniques for tracking projected fiducial markers that move and overlap. We introduce a range of example applications that demonstrate the applicability of our system to real-world scenarios such as mobile content exchange, gaming, and education.

Spatial Sketch is a three-dimensional (3D) sketch application that bridges between physical movement and the fabrication of objects in the real world via cut planar materials. This paper explores the rationale and details behind the development of the Spatial Sketch application, and presents our observations from user testing and a hands-on lamp shade design workshop. Finally we reflect upon the relevance of embodied forms of human computer interaction for use in digital fabrication.

This paper presents an overview of Alchemy, an experimental drawing application aimed at exploring how we can sketch, draw, and create on computers in new ways. Alchemy focuses on the absolute initial stage of the creative process, to provide an expanded range of possibilities for serendipitous sketching and shape creation. The main aim of Alchemy is to explore how computer based forms of drawing can extend the early stage idea creation process.

This paper tracks the author's current research and art practice, focused on the production of systems for artistic creation and examining how such systems contribute to an engaging interactive experience.

This paper tracks the development of the author's work entitled Light Tracer, and examines the surrounding issues of user authorship and creativity within interactivity.Light Tracer is an interactive system which invites the participant to write, draw and trace images in real physical space. The participant is situated in front of a screen reflecting their own image, and by manipulating a series of light sources, marks can be left onscreen such as drawings, messages, traces of physical objects such as faces, hands and bodies.It is the argument of the author that by allowing the user an optimum level of creative authorship within an interactive work, the user can be successfully engaged with the experience of the interaction and in turn produce and create themselves.