Deep forecasting models show output heads significantly improve performance on fat-tailed financial returns.
problem Improving deep learning models for forecasting fat-tailed financial returns.
method Comparison of backbone architectures and output heads (point, Gaussian, Gaussian mixture) on S&P 500 monthly log-returns.
result Switching from point to Gaussian heads improves CRPS by about 1.3 percent, and from Gaussian to mixture adds another 2.4 percent.
MTL-NAS combines NAS with GP-MTL for task-agnostic multi-task learning.
problem Designing architectures for diverse tasks with varying priors.
method Disentangled GP-MTL networks, hierarchical feature sharing, and gradient-based search.
result General-purpose model trained once can adapt to multiple tasks.
Generative model designs highly designable proteins using geometric algebra.
problem Creating proteins with diverse and statistically accurate secondary structures.
method Introduced a geometric algebra flow matching model (FrameFlow) with Clifford Frame Attention (CFA) for protein backbone design.
result Achieved high designability, diversity, and novelty in protein backbone sampling.
V-HMN integrates memory mechanisms for improved image recognition.
problem Limited interpretability and high data requirements of existing vision backbones.
method Brain-inspired hierarchical memory modules with iterative refinement.
result V-HMN achieves strong performance on image classification benchmarks.
RPN 2 improves function learning by modeling data interdependence.
problem Invalid assumption of input data independence leads to performance degradation.
method Integrates data and structural interdependence functions into RPN 2's architecture.
result Significantly improves learning performance and expands unifying potential.
BiTAT improves neural network quantization for edge devices by focusing on weight dependencies and disentangling them.
problem Performance degradation of compact neural networks under extreme quantization.
method Task-dependent Aggregated Transformation (BiTAT) method that orthonormalizes weights and progressively quantizes them.
result BiTAT effectively preserves model performance on ImageNet and CIFAR-100 with compact backbones.
We detect the backbone of the weighted bipartite network of the Japanese credit market relationships. The backbone is detected by adapting a general method used in the investigation of weighted networks. With this approach we detect a backbone that is statistically validated against a null hypothesis of uniform diversi…
FiGS searches over a larger space of architectures for efficient mobile models.
problem Designing small, efficient deep networks for mobile devices.
method Differentiable search method using sparse regularization and Logistic-Sigmoid distribution.
result FiGS produces state-of-the-art parameter-efficient models on ImageNet and improves object detection performance.
Frame Averaging makes neural networks invariant or equivariant to new symmetries.
problem Designing neural networks that respect symmetries while being expressive and efficient.
method Introduces Frame Averaging (FA) as a systematic framework to adapt architectures to become invariant or equivariant to new symmetries.
result Frame Averaging guarantees exact invariance or equivariance while being simpler to compute than full group averaging.
GAPA method provides efficient uncertainty quantification for pretrained networks.
problem Reliable uncertainty estimates for pretrained models are challenging.
method Post-hoc Gaussian Process Activations (GAPA) method that shifts Bayesian modeling from weights to activations.
result GAPA method provides efficient uncertainty quantification without altering the backbone's predictions.
Researchers use shape analysis to recover protein structures from Cryo-EM data.
problem Recovering the three-dimensional backbone structure of single polypeptide proteins from noisy tomographic projections.
method Shape analysis and matrix Lie group actions to deform point clouds to match 2D tomography data.
result Optimal deformations are computed to recover the three-dimensional backbone structure of proteins.
A new method scales sparse machine learning to ultra-high dimensional problems.
problem Sparse and interpretable machine learning in ultra-high dimensional data.
method Two-phase approach: backbone set determination followed by reduced problem solving.
result The backbone set contains truly relevant features with high probability.
Image segmentation is widely used in a variety of computer vision tasks, such as object localization and recognition, boundary detection, and medical imaging. This thesis proposes deep learning architectures to improve automatic object localization and boundary delineation for salient object segmentation in natural ima…
LLoCa makes any network Lorentz-equivariant, achieving high accuracy and efficiency.
problem Limitations of specialized layers in Lorentz-equivariant neural networks.
method LLoCa framework using local reference frames and geometric message passing.
result Models achieve competitive and state-of-the-art accuracy on particle physics tasks.
A new diffusion model generates novel protein backbones without relying on pretrained networks.
problem Generating novel protein backbones without relying on pretrained networks.
method Developed a SE(3) invariant diffusion model on multiple frames, called FrameDiff.
result Generated designable protein monomers up to 500 amino acids without pretrained networks.
New kernel connects deep learning to optimization, improving generalization.
problem Understanding why deep neural networks generalize well at over-parameterization.
method Established Neural Optimization Kernel (NOK) linking deep NN to optimization problems.
result New generalization bound for deep structured approximated NOK architecture.
In this paper, we investigate the statistical features of the weighted international-trade network. By finding the maximum weight spanning trees for this network we make the extraction of the truly relevant connections forming the network's backbone. We discuss the role of large-sized countries (strongest economies) in…
This paper evaluates heuristics and hyperparameters in weight-sharing NAS methods.
problem Improving the performance of weight-sharing NAS methods.
method Systematic evaluation of heuristics and hyperparameters in weight-sharing NAS algorithms.
result Some heuristics negatively impact super-net and stand-alone performance correlation.
Neural Networks (NN) have recently emerged as backbone of several sensitive applications like automobile, medical image, security, etc. NNs inherently offer Partial Fault Tolerance (PFT) in their architecture; however, the biased PFT of NNs can lead to severe consequences in applications like cryptography and security …
Hydra boosts efficiency for long-context reasoning in resource-constrained settings.
problem Quadratic complexity of transformers limits long-context reasoning in resource-constrained systems.
method Hydra uses a modular architecture with adaptive routing between sparse global attention, mixture-of-experts, and dual memories.
result Hydra achieves significant throughput and accuracy improvements for long-context reasoning.
Class-conditional generative models are crucial tools for data generation from user-specified class labels. Existing approaches for class-conditional generative models require nontrivial modifications of backbone generative architectures to model conditional information fed into the model. This paper introduces a plug-…
GenSBI offers JAX-native SBI methods for natural sciences.
problem Lack of native SBI libraries in JAX for natural sciences.
method Flow and diffusion generative models in JAX.
result Near-ideal mean C2ST scores on SBIBM tasks.
One of the greatest challenges in the design of a real-time perception system for autonomous driving vehicles and drones is the conflicting requirement of safety (high prediction accuracy) and efficiency. Traditional approaches use a single frame rate for the entire system. Motivated by the observation that the lack of…
Correlation Networks (CNs) inherently suffer from redundant information in their network topology. Bayesian Networks (BNs), on the other hand, include only non-redundant information (from a probabilistic perspective) resulting in a sparse topology from which generalizable physical features can be extracted. We advocate…
New approach uses contrastive learning for better wireless power control.
problem Improving power control in wireless networks.
method A multi-layer perceptron with a contrastive learning backbone and head.
result Significant gains in sum-throughput and sample efficiency over supervised learning.
We propose Mish, a novel self-regularized non-monotonic activation function which can be mathematically defined as: f(x)=xtanh(softplus(x)). As activation functions play a crucial role in the performance and training dynamics in neural networks, we validated experimentally on several well-known benchmarks…
Aux-NAS uses auxiliary labels to improve primary task performance without extra inference cost.
problem Improving primary task performance using auxiliary labels without increasing inference cost.
method Architecture-based approach with a flexible asymmetric structure for primary and auxiliary tasks, using Neural Architecture Search (NAS) to evolve networks with only primary-to-auxiliary connections.
result Achieves improved performance on multiple tasks without increasing inference cost.
Mapper-GIN simplifies 3D point cloud classification with lightweight structure.
problem Robust 3D point cloud classification under corruption.
method Mapper algorithm for structural decomposition, GIN for graph classification.
result Mapper-GIN achieves competitive accuracy with minimal parameters.
New method learns diverse protein scaffolds for motif design.
problem Designing long, diverse protein scaffolds for specific motifs.
method E(3)-equivariant graph neural network for diffusion modeling.
result First to guarantee conditional sampling from diffusion models.
LeJEPA provides a scalable, theory-driven approach to self-supervised learning.
problem Lack of practical guidance and theory in JEPAs.
method Identified optimal Gaussian distribution and introduced SIGReg objective.
result LeJEPA achieves state-of-the-art performance with minimal hyperparameters and heuristics.
Few-step protein backbone generators reduce sampling time by over 20x.
problem Computational bottleneck in diffusion-based protein generation models.
method Score distillation adapted for protein backbone generation, combined with inference time noise modulation.
result Significant reduction in sampling time (20+ fold) while maintaining comparable performance.
Paper improves few-shot classification accuracy using feature distribution preprocessing.
problem Challenges of few-shot classification due to limited labelled samples.
method Proposes a novel transfer-based method that preprocesses feature vectors to Gaussian-like distributions and uses optimal-transport inspired algorithms.
result Achieves state-of-the-art accuracy on standardized vision benchmarks.
Multi-headed ensembles boost model performance with faster training.
problem Limited computational resources hinder ensemble search performance.
method Extend NES to multi-headed ensembles, leveraging end-to-end training and one-shot NAS methods.
result Multi-headed ensemble search finds robust ensembles 3 times faster with comparable performance.
Proteins are the major building blocks of life, and actuators of almost all chemical and biophysical events in living organisms. Their native structures in turn enable their biological functions which have a fundamental role in drug design. This motivates predicting the structure of a protein from its sequence of amino…
Kolmogorov-Arnold Networks promise scalable performance in high dimensions.
problem Curse of dimensionality in multilayer perceptrons.
method Kolmogorov-Arnold representation theorem and interpolation methods.
result Kolmogorov-Arnold Networks achieve true freedom from the curse of dimensionality.
ModSSC unifies semi-supervised classification for various data types.
problem Fragmented support for semi-supervised classification across different methods, settings, and data types.
method ModSSC is a modular Python framework that supports reproducible and controlled experimentation for semi-supervised classification on heterogeneous data.
result ModSSC enables systematic comparison of semi-supervised learning across various datasets and model backbones.
Paper tackles fairness issues in GNNs by proposing ELEGANT for certification.
problem Fairness issues in GNN predictions due to graph data perturbations.
method Proposes ELEGANT framework for certifying fairness of any GNN without assumptions or re-training.
result The fairness of any GNN backbone is impossible to be corrupted under certain perturbation budgets.
AbDiffuser generates full-atom antibodies with sequence and structure fidelity.
problem Generating high-fidelity antibodies with both structure and sequence information.
method Equivariant and physics-informed diffusion model with novel protein structure representation.
result AbDiffuser generates antibodies with sequence and structural properties matching a reference set.
An analysis of the Japanese credit market in 2004 between banks and quoted firms is done in this paper using the tools of the networks theory. It can be pointed out that: (i) a backbone of the credit channel emerges, where some links play a crucial role; (ii) big banks privilege long-term contracts; the "minimal spanni…
U-Cast simplifies AI weather forecasting with a standard U-Net and efficient training.
problem Complex AI models limit accessibility and cost for weather forecasting.
method Simple U-Net backbone, deterministic pre-training, and probabilistic fine-tuning with Monte Carlo Dropout.
result U-Cast matches or exceeds state-of-the-art models in accuracy while reducing training and inference costs.
Paper introduces new loss functions for Siamese networks using FDA.
problem Training Siamese networks with improved loss functions.
method Proposes Fisher Discriminant Triplet (FDT) and Fisher Discriminant Contrastive (FDC) loss functions based on FDA.
result Shows effectiveness of FDT and FDC on MNIST and histopathology datasets.
A new pruning method improves neural network efficiency and accuracy.
problem Optimizing neural network efficiency and accuracy through pruning.
method Formulated as a Knapsack Problem, the method optimizes trade-off between neuron importance and computational cost. Channels are pruned while maintaining network structure, and fine-tuned using inner knowledge distillation from parent network.
result State-of-the-art pruning results on ImageNet, CIFAR-10, and CIFAR-100.
SPIRE enables efficient federated learning for diffusion models by separating client-specific embeddings from a shared backbone.
problem Large diffusion models are impractical for federated learning due to their size.
method SPIRE separates the network into a global backbone and client-specific embeddings, enabling efficient finetuning.
result SPIRE achieves parameter-efficient finetuning, updating only a small fraction of weights.
Enhanced image denoising with MWRDCNN using residual dense blocks.
problem Image denoising with improved performance and robustness.
method Multi-wavelet residual dense convolutional neural network (MWRDCNN) with residual dense blocks (RDBs).
result Significantly improved performance in image denoising compared to existing techniques.
PoET-BiN reduces power consumption in neural networks on embedded devices.
problem Power inefficiency in neural network implementations on embedded platforms.
method Look-Up Table based implementation with a modified Decision Tree approach.
result Near state-of-the-art results with up to 6 orders of magnitude energy reduction.
BackboneLearn speeds up MIO-based machine learning problems.
problem Scaling mixed-integer optimization problems in machine learning.
method An open-source Python framework for MIO problems with indicator variables.
result Solves MIO problems faster and more accurately than existing methods.
Structural and topological information play a key role in modeling flow and transport through fractured rock in the subsurface. Discrete fracture network (DFN) computational suites such as dfnWorks are designed to simulate flow and transport in such porous media. Flow and transport calculations reveal that a small back…
The backbone of most proteins forms an open curve. To study their entanglement, a common strategy consists in searching for the presence of knots in their backbones using topological invariants. However, this approach requires to close the curve into a loop, which alters the geometry of curve. Knoto-ID allows evaluatin…