While 2D generative adversarial networks have enabled high-resolution image synthesis, they largely lack an understanding of the 3D world and the image formation process. Thus, they do not provide precise control over camera viewpoint or object pose. To address this problem, several recent approaches leverage intermediate voxel-based representations in combination with differentiable rendering. However, existing methods either produce low image resolution or fall short in disentangling camera and scene properties, e.g., the object identity may vary with the viewpoint. In this paper, we propose a generative model for radiance fields which have recently proven successful for novel view synthesis of a single scene. In contrast to voxel-based representations, radiance fields are not confined to a coarse discretization of the 3D space, yet allow for disentangling camera and scene properties while degrading gracefully in the presence of reconstruction ambiguity. By introducing a multi-scale patch-based discriminator, we demonstrate synthesis of high-resolution images while training our model from unposed 2D images alone. We systematically analyze our approach on several challenging synthetic and real-world datasets. Our experiments reveal that radiance fields are a powerful representation for generative image synthesis, leading to 3D consistent models that render with high fidelity.

The ambiguity function is fundamental to radar waveform design, characterizing range and Doppler resolution capabilities. However, its traditional formulation involves non-differentiable operations, preventing integration with gradient-based optimization methods and modern machine learning frameworks. This paper presents the first complete mathematical framework and computational implementation for differentiable radar ambiguity functions. Our approach addresses the fundamental technical challenges that have prevented the radar community from leveraging automatic differentiation: proper handling of complex-valued gradients using Wirtinger calculus, efficient computation through parallelized FFT operations, numerical stability throughout cascaded operations, and composability with arbitrary differentiable operations. We term this approach GRAF (Gradient-based Radar Ambiguity Functions), which reformulates the ambiguity function computation to maintain mathematical equivalence while enabling gradient flow through the entire pipeline. The resulting implementation provides a general-purpose differentiable ambiguity function compatible with modern automatic differentiation frameworks, enabling new research directions including neural network-based waveform generation with ambiguity constraints, end-to-end optimization of radar systems, and integration of classical radar theory with modern deep learning. We provide complete implementation details and demonstrate computational efficiency suitable for practical applications. This work establishes the mathematical and computational foundation for applying modern machine learning techniques to radar waveform design, bridging classical radar signal processing with automatic differentiation frameworks.

In general, the reviewers were positive about the paper: the proposed GRAF can successfully capture high-res 3D-aware image synthesis from unposed images, the patch-based discriminator is effective and the paper is well written, while there was concern that the paper felt like a combination of HoloGAN and NeRF. After the rebuttal and discussion, the reviewers all voted for acceptance. Please put the important points in the rebuttal to the final version.

Weaknesses: The novelty is slightly limited, as this paper mostly feels like a combination of HoloGAN and NeRF. However, I think that this is not a deal-breaker, and I appreciate the authors' efforts to distill the experiments into key takeaways that can be relevant for future progress in 3D-aware generative modeling. I am surprised about the ray sampling used in the patch discriminator. Since the same discriminator views different decimated versions of the image, I am surprised that the effects of aliasing do not prevent learning high-frequency details (I would be less concerned if different discriminators were used for different levels of subsampling, but in the proposed method, the same discriminator sees differently aliased versions of the same content). Maybe this would be relevant when trying to scale up to even higher resolutions?

While 2D generative adversarial networks have enabled high-resolution image synthesis, they largely lack an understanding of the 3D world and the image formation process. Thus, they do not provide precise control over camera viewpoint or object pose. To address this problem, several recent approaches leverage intermediate voxel-based representations in combination with differentiable rendering. However, existing methods either produce low image resolution or fall short in disentangling camera and scene properties, e.g., the object identity may vary with the viewpoint. In this paper, we propose a generative model for radiance fields which have recently proven successful for novel view synthesis of a single scene.

In this paper, we propose a novel activation function tailored specifically for graph data in Graph Neural Networks (GNNs). Motivated by the need for graph-adaptive and flexible activation functions, we introduce DiGRAF, leveraging Continuous Piecewise-Affine Based (CPAB) transformations, which we augment with an additional GNN to learn a graph-adaptive diffeomorphic activation function in an end-to-end manner. In addition to its graph-adaptivity and flexibility, DiGRAF also possesses properties that are widely recognized as desirable for activation functions, such as differentiability, boundness within the domain and computational efficiency. We conduct an extensive set of experiments across diverse datasets and tasks, demonstrating a consistent and superior performance of DiGRAF compared to traditional and graph-specific activation functions, highlighting its effectiveness as an activation function for GNNs.

Performance prediction has been a key part of the neural architecture search (NAS) process, allowing to speed up NAS algorithms by avoiding resource-consuming network training. Although many performance predictors correlate well with ground truth performance, they require training data in the form of trained networks. Recently, zero-cost proxies have been proposed as an efficient method to estimate network performance without any training. However, they are still poorly understood, exhibit biases with network properties, and their performance is limited. Inspired by the drawbacks of zero-cost proxies, we propose neural graph features (GRAF), simple to compute properties of architectural graphs. GRAF offers fast and interpretable performance prediction while outperforming zero-cost proxies and other common encodings. In combination with other zero-cost proxies, GRAF outperforms most existing performance predictors at a fraction of the cost.

A large number of real-world networks include multiple types of nodes and edges. Graph Neural Network (GNN) emerged as a deep learning framework to generate node and graph embeddings for downstream machine learning tasks. However, popular GNN-based architectures operate on single homogeneous networks. Enabling them to work on multiple networks brings additional challenges due to the heterogeneity of the networks and the multiplicity of the existing associations. In this study, we present a computational approach named GRAF (Graph Attention-aware Fusion Networks) utilizing GNN-based approaches on multiple networks with the help of attention mechanisms and network fusion. Using attention-based neighborhood aggregation, GRAF learns the importance of each neighbor per node (called node-level attention) followed by the importance of association (called association-level attention). Then, GRAF processes a network fusion step weighing each edge according to learned node- and association-level attentions. Considering that the fused network could be a highly dense network with many weak edges depending on the given input networks, we included an edge elimination step with respect to edges' weights. Finally, GRAF utilizes Graph Convolutional Network (GCN) on the fused network and incorporates node features on graph-structured data for a node classification or a similar downstream task. To demonstrate GRAF's generalizability, we applied it to four datasets from different domains and observed that GRAF outperformed or was on par with the baselines, state-of-the-art methods, and its own variations for each node classification task. Source code for our tool is publicly available at https://github.com/bozdaglab/GRAF .