Immutability is common in the programming mainstream: deep immutability is the default in functional languages while imperative languages typically provide opt-in support for shallow immutability, usually enforced through static checking. Python is a dynamic imperative language where mutability is inherent: not only are most objects mutable, but programs themselves---modules, classes, functions---are represented by mutable objects at run-time, and libraries routinely rely on this mutability. This makes adding immutability to Python a significant challenge. This paper presents the design and implementation of deep immutability for Python. Our primary motivation is to permit multiple sub-interpreters to directly share object references, which currently requires costly serialisation. Sharing via immutability introduces a soundness challenge, as a violation could corrupt the interpreter's state. We identify numerous challenges that stem from decades of design decisions that did not anticipate immutability, and show how they can be overcome through two complementary techniques: detachment, which severs run-time links that would cause immutability to propagate too widely, and freezability, which gives objects run-time control over whether and how they may become immutable. Together, these principles form a general design pattern for deep immutability in dynamic languages. We validate our design with an implementation on CPython 3.15 that is backwards-compatible with existing programs and enables direct, zero-copy sharing of immutable objects across sub-interpreters.

The ways in which the components of a program interact with each other in a concurrent setting can be considerably more complex than in a sequential setting. The core problem is unrestricted shared mutable state. An alternative to unrestricted shared mutable state is to restrict the sharing using Ownership. Ownership can turn what would have been a race into a deterministic failure that can be explained to the programmer. However, Ownership has predominantly taken place in statically typed languages. In this paper, we explore retrofitting an existing dynamically typed programming language with an ownership model based on regions. Our core aim is to provide safe concurrency, that is, the ownership model should provide deterministic dynamic failures of ownership that can be explained to the programmer. We present a dynamic model of ownership that provides ownership of groups objects called regions. We provide dynamic enforcement of our region discipline, which we have implemented in a simple interpreter that provides a Pythonlike syntax and semantics, and report on our first steps into integrating it into an existing language, Python.

Compilation is all about tree rewriting. In functional languages where all data is tree-shaped, tree rewriting is facilitated by pattern matching, but data immutability leads to copying for each update. In object-oriented languages like Java or C++, a standard approach is to use the visitor pattern, which increases modularization but also adds indirection and introduces boilerplate code. In this paper, we introduce Trieste -- a novel tree-rewriting DSL, combining the power of C++ with the expressivity of pattern matching. In Trieste, sequences of rewrite passes can be used to read a file to produce an abstract syntax tree (AST), convert from one AST to another, or write an AST to disk. Each pass rewrites an AST in place using subtree pattern matching, where the result is dynamically checked for well-formedness. Checking the well-formedness of trees dynamically enables flexibly changing the tree structure without having to define new data types for each intermediate representation. The well-formedness specification can also be used for scoped name binding and generating random well-formed trees for fuzz testing in addition to checking the shape of trees. Trieste has been used to build fully compliant parsers for YAML and JSON, a transpiler from YAML to JSON, and a compiler and interpreter for the policy language Rego.

Abstract The recent increase in popularity of volumetric representations for scene reconstruction and novel view synthesis has put renewed focus on animating volumetric content at high visual quality and in real‐time. While implicit deformation methods based on learned functions can produce impressive results, they are ‘black boxes’ to artists and content creators, they require large amounts of training data to generalize meaningfully, and they do not produce realistic extrapolations outside of this data. In this work, we solve these issues by introducing a volume deformation method which is real‐time even for complex deformations, easy to edit with off‐the‐shelf software and can extrapolate convincingly. To demonstrate the versatility of our method, we apply it in two scenarios: physics‐based object deformation and telepresence where avatars are controlled using blendshapes. We also perform thorough experiments showing that our method compares favourably to both volumetric approaches combined with implicit deformation and methods based on mesh deformation.

Recent work on Neural Radiance Fields (NeRF) showed how neural networks can be used to encode complex 3D environments that can be rendered photorealistically from novel viewpoints. Rendering these images is very computationally demanding and recent improvements are still a long way from enabling interactive rates, even on high-end hardware. Motivated by scenarios on mobile and mixed reality devices, we propose FastNeRF, the first NeRF-based system capable of rendering high fidelity photorealistic images at 200Hz on a high-end consumer GPU. The core of our method is a graphics-inspired factorization that allows for (i) compactly caching a deep radiance map at each position in space, (ii) efficiently querying that map using ray directions to estimate the pixel values in the rendered image. Extensive experiments show that the proposed method is 3000 times faster than the original NeRF algorithm and at least an order of magnitude faster than existing work on accelerating NeRF, while maintaining visual quality and extensibility.

Decision trees and randomized forests are widely used in computer vision and machine learning. Standard algorithms for decision tree induction optimize the split functions one node at a time according to some splitting criteria. This greedy procedure often leads to suboptimal trees. In this paper, we present an algorithm for optimizing the split functions at all levels of the tree jointly with the leaf parameters, based on a global objective. We show that the problem of finding optimal linear-combination (oblique) splits for decision trees is related to structured prediction with latent variables, and we formulate a convex-concave upper bound on the tree's empirical loss. Computing the gradient of the proposed surrogate objective with respect to each training exemplar is O(d^2), where d is the tree depth, and thus training deep trees is feasible. The use of stochastic gradient descent for optimization enables effective training with large datasets. Experiments on several classification benchmarks demonstrate that the resulting non-greedy decision trees outperform greedy decision tree baselines.

We propose semantic texton forests, efficient and powerful new low-level features. These are ensembles of decision trees that act directly on image pixels, and therefore do not need the expensive computation of filter-bank responses or local descriptors. They are extremely fast to both train and test, especially compared with k-means clustering and nearest-neighbor assignment of feature descriptors. The nodes in the trees provide (i) an implicit hierarchical clustering into semantic textons, and (ii) an explicit local classification estimate. Our second contribution, the bag of semantic textons, combines a histogram of semantic textons over an image region with a region prior category distribution. The bag of semantic textons is computed over the whole image for categorization, and over local rectangular regions for segmentation. Including both histogram and region prior allows our segmentation algorithm to exploit both textural and semantic context. Our third contribution is an image-level prior for segmentation that emphasizes those categories that the automatic categorization believes to be present. We evaluate on two datasets including the very challenging VOC 2007 segmentation dataset. Our results significantly advance the state-of-the-art in segmentation accuracy, and furthermore, our use of efficient decision forests gives at least a five-fold increase in execution speed.

Abstract Composite images are synthesized from existing photographs by artists who make concept art, e.g., storyboards for movies or architectural planning. Current techniques allow an artist to fabricate such an image by digitally splicing parts of stock photographs. While these images serve mainly to “quickly”convey how a scene should look, their production is laborious. We propose a technique that allows a person to design a new photograph with substantially less effort. This paper presents a method that generates a composite image when a user types in nouns, such as “boat”and “sand.”The artist can optionally design an intended image by specifying other constraints. Our algorithm formulates the constraints as queries to search an automatically annotated image database. The desired photograph, not a collage, is then synthesized using graph‐cut optimization, optionally allowing for further user interaction to edit or choose among alternative generated photos. An implementation of our approach, shown in the associated video, demonstrates our contributions of (1) a method for creating specific images with minimal human effort, and (2) a combined algorithm for automatically building an image library with semantic annotations from any photo collection.

The majority of machine learning systems for object recognition is limited by their requirement of single labelled images for training, which are difficult to create or obtain in quantity. It is therefore impractical to use methods or techniques which require such data to build object recognizers for more than a relatively small subset of object classes. Instead, far more abundant multilabel data provides a ready means to create object recognition systems which are able to deal with large numbers of classes. In this paper we present a new object recognition system named MLBoost which learns from multi-label data through boosting and improves on state-of-the-art multi-label annotation and labelling systems. The system is trained on images with accompanying text and at no time is told which parts of each image correspond to which words, and as such the process is unsupervised. Having once been trained it is able to give segment labels and a list of descriptive words (an annotation) for any novel image.