SEARCH · Search arXiv

Results for “transformer architecture”

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,027 records · Page 57Linked to original sources

arXiv2023

SiamixFormer: a fully-transformer Siamese network with temporal Fusion for accurate building detection and change detection in bi-temporal remote sensing images

Building detection and change detection using remote sensing images can help urban and rescue planning. Moreover, they can be used for building damage assessment after natural disasters. Currently, most of the existing models for building detection use only one image (pre-disaster image) to detect buildings. This is based on the idea that post-disaster images reduce the model's performance because of presence of destroyed buildings. In this paper, we propose a siamese model, called SiamixFormer, which uses pre- and post-disaster images as input. Our model has two encoders and has a hierarchical transformer architecture. The output of each stage in both encoders is given to a temporal transformer for feature fusion in a way that query is generated from pre-disaster images and (key, value) is generated from post-disaster images. To this end, temporal features are also considered in feature fusion. Another advantage of using temporal transformers in feature fusion is that they can better maintain large receptive fields generated by transformer encoders compared with CNNs. Finally, the output of the temporal transformer is given to a simple MLP decoder at each stage. The SiamixFormer model is evaluated on xBD, and WHU datasets, for building detection and on LEVIR-CD and CDD datasets for change detection and could outperform the state-of-the-art.

cs.CV↗

arXiv2025

Folded Context Condensation in Path Integral Formalism for Infinite Context Transformers

In this work, we present a generalized formulation of the Transformer algorithm by reinterpreting its core mechanisms within the framework of Path Integral formalism. In this perspective, the attention mechanism is recast as a process that integrates all possible transition paths leading to future token states, with temporal evolution governed by the Feed-Forward Network. By systematically mapping each component of the Transformer to its counterpart in the Path Integral formulation, we obtain a more compact and efficient representation, in which the contextual information of a sequence is condensed into memory-like segments. These segments are recurrently processed across Transformer layers, enabling more effective long-term information retention. We validate the effectiveness of this approach through the Passkey retrieval task and a summarization task, demonstrating that the proposed method preserves historical information while exhibiting memory usage that scales linearly with sequence length. This contrasts with the non-linear memory growth typically observed in standard attention mechanisms. We expect that this quantum-inspired generalization of the Transformer architecture will open new avenues for enhancing both the efficiency and expressiveness of future Transformer models.

hep-ph↗

arXiv2026

From Tokens to States: LLMs as a Special Case of World Models and the Continuous Path Beyond

The AI community has framed the relationship between large language models (LLMs) and world models as a dichotomy: LLMs predict tokens; world models simulate reality. Yann LeCun argues in 2022 that reaching general intelligence requires abandoning autoregressive token prediction in favour of latent-space architectures. This framing is unnecessarily binary. Two claims will be defended. First, LLMs are a degenerate special case of world models: the state space is the set of all token sequences, the only action is appending one token, and world models are therefore a strict generalisation of LLMs, not a replacement. Second, there is a natural continuous spectrum from NTP to JEPA, with multi-token prediction, future-summary prediction, and next-latent prediction as intermediate stations already populated by current research. Moving along this spectrum relaxes the LLM constraints one by one. It also progressively surrenders the two practical advantages that make LLMs trainable at scale: internet-scale self-supervised data, and a transformer architecture co-designed for discrete token prediction. Both are examined as open research questions: the data question (the cliff from self-supervised text to instrumented action-labelled environments) and the architecture question (whether the transformer generalises to continuous-state prediction, or whether a new primitive is needed).

cs.CL↗

arXiv2026

Can Transformers Really Do It All? On the Compatibility of Inductive Biases Across Tasks

Transformers are remarkably versatile and their design is largely consistent across a variety of applications. But are they optimal for any given task or dataset? The answer may be key for pushing AI beyond merely scaling current designs. *Method.* We present a method to optimize a transformer architecture for a given dataset, which we use as a tool to study optimal task-specific inductive biases. This method replaces the most important non-linearities (GeLUs,;softmax) with functions learned on held-out data. We then train the resulting architectures on other datasets, as a way to evaluate the compatibility between pairs of tasks. Findings. On algorithmic toy tasks, we identify new architectures with dramatic improvements in learning speed, in- and out-of-distribution generalization, and stability across seeds. The new designs prove very task-specific however, and indicate that these tasks require inductive biases very different from those of standard transformers. On code and language modeling datasets, we also find architectures with consistent, yet smaller improvements. These designs transfer much better across datasets and domains (English & computer code). Implications. Our results show that standard transformers are rarely a local optimum in the space of architectures. Simple alternatives can perform much better but sacrifice universality. This suggests that there may be room for improved architectures that better support multiple capabilities simultaneously, such as fluency and robust reasoning.

cs.LG↗

arXiv2026

OrthoFormer: Instrumental Variable Estimation in Transformer Hidden States via Neural Control Functions

Transformer architectures excel at sequential modeling yet remain fundamentally limited by correlational learning - they capture spurious associations induced by latent confounders rather than invariant causal mechanisms. We identify this as an epistemological challenge: standard Transformers conflate static background factors (intrinsic identity, style, context) with dynamic causal flows (state evolution, mechanism), leading to catastrophic out-of-distribution failure. We propose OrthoFormer, a causally grounded architecture that embeds instrumental variable estimation directly into Transformer blocks via neural control functions. Our framework rests on four theoretical pillars: Structural Directionality (time-arrow enforcement), Representation Orthogonality (latent-noise separation), Causal Sparsity (Markov Blanket approximation), and End-to-End Consistency (gradient- detached stage separation). We prove that OrthoFormer achieves bias strictly less than OLS for any valid instrument lag, with residual bias decaying geometrically as O(\r{ho}k ). We characterize the bias-variance-exogeneity trilemma inherent in self-instrumenting and identify the neural forbidden regression - where removing gradient detachment improves prediction loss while destroying causal validity. Experiments confirm all theoretical predictions. OrthoFormer represents a paradigm shift from correlational to causal sequence modeling, with implications for robustness, interpretability, and reliable decision-making under distribution shift.

cs.LG↗

arXiv2026

Geometric and Dynamic Scaling in Deep Transformers

Despite their empirical success, pushing Transformer architectures to extreme depth often leads to a paradoxical failure: representations become increasingly redundant, lose rank, and ultimately collapse. Existing explanations largely attribute this phenomenon to optimization instability or vanishing gradients, yet such accounts fail to explain why collapse persists even under modern normalization and initialization schemes. In this paper, we argue that the collapse of deep Transformers is fundamentally a geometric problem. Standard residual updates implicitly assume that feature accumulation is always beneficial, but offer no mechanism to constrain update directions or to erase outdated information. As depth increases, this leads to systematic drift off the semantic manifold and monotonic feature accumulation, causing representational degeneracy. We propose a unified geometric framework that addresses these failures through two orthogonal principles. First, manifold-constrained hyper-connections restrict residual updates to valid local tangent directions, preventing uncontrolled manifold drift. Second, deep delta learning introduces data-dependent, non-monotonic updates that enable reflection and erasure of redundant features rather than their unconditional accumulation. Together, these mechanisms decouple the direction and sign of feature updates, yielding a stable geometric evolution across depth. We term the resulting architecture the Manifold-Geometric Transformer (MGT). Our analysis predicts that enforcing geometric validity while allowing dynamic erasure is essential for avoiding rank collapse in ultra-deep networks. We outline an evaluation protocol for Transformers exceeding 100 layers to test the hypothesis that geometry, rather than depth itself, is the key limiting factor in deep representation learning.

cs.LG↗

arXiv2025

X-ray illicit object detection using hybrid CNN-transformer neural network architectures

In the field of X-ray security applications, even the smallest details can significantly impact outcomes. Objects that are heavily occluded or intentionally concealed pose a great challenge for detection, whether by human observation or through advanced technological applications. While certain Deep Learning (DL) architectures demonstrate strong performance in processing local information, such as Convolutional Neural Networks (CNNs), others excel in handling distant information, e.g., transformers. In X-ray security imaging the literature has been dominated by the use of CNN-based methods, while the integration of the two aforementioned leading architectures has not been sufficiently explored. In this paper, various hybrid CNN-transformer architectures are evaluated against a common CNN object detection baseline, namely YOLOv8. In particular, a CNN (HGNetV2) and a hybrid CNN-transformer (Next-ViT-S) backbone are combined with different CNN/transformer detection heads (YOLOv8 and RT-DETR). The resulting architectures are comparatively evaluated on three challenging public X-ray inspection datasets, namely EDS, HiXray, and PIDray. Interestingly, while the YOLOv8 detector with its default backbone (CSP-DarkNet53) is generally shown to be advantageous on the HiXray and PIDray datasets, when a domain distribution shift is incorporated in the X-ray images (as happens in the EDS datasets), hybrid CNN-transformer architectures exhibit increased robustness. Detailed comparative evaluation results, including object-level detection performance and object-size error analysis, demonstrate the strengths and weaknesses of each architectural combination and suggest guidelines for future research. The source code and network weights of the models employed in this study are available at https://github.com/jgenc/xray-comparative-evaluation.

cs.CV↗

arXiv2023

HCT: Hybrid Convnet-Transformer for Parkinson's disease detection and severity prediction from gait

In this paper, we propose a novel deep learning method based on a new Hybrid ConvNet-Transformer architecture to detect and stage Parkinson's disease (PD) from gait data. We adopt a two-step approach by dividing the problem into two sub-problems. Our Hybrid ConvNet-Transformer model first distinguishes healthy versus parkinsonian patients. If the patient is parkinsonian, a multi-class Hybrid ConvNet-Transformer model determines the Hoehn and Yahr (H&Y) score to assess the PD severity stage. Our hybrid architecture exploits the strengths of both Convolutional Neural Networks (ConvNets) and Transformers to accurately detect PD and determine the severity stage. In particular, we take advantage of ConvNets to capture local patterns and correlations in the data, while we exploit Transformers for handling long-term dependencies in the input signal. We show that our hybrid method achieves superior performance when compared to other state-of-the-art methods, with a PD detection accuracy of 97% and a severity staging accuracy of 87%. Our source code is available at: https://github.com/SafwenNaimi

cs.CV↗

arXiv2026

Denoising Diffusion Generative Models Secretly Calculate Attentions

Denoising diffusion models are the dominant architecture for image generation, whereas most natural language generation and modeling are primarily handled by well-known transformer architectures employing attention mechanism. Here, we show that diffusion models also inherently use an attention mechanism very similar to that of transformers. Therefore, attention emerges as a universal machine learning principle, based on a general training objective. We also show similarities in basic functional principle of auto-encoders and attention-based models. These equivalences allows us to interchange these designs based on practical requirements. As an example, we can reformulate the diffusion framework to reduce the lengthy training process and computation-intensive image generation. Using this approach, a simplified algorithm is proposed for image generation which is based on attention mechanism. Results show that the attention-based implementation achieves comparable performance with significantly less effort and computational resources.

cs.AI↗

arXiv2025

COSMO: Combination of Selective Memorization for Low-cost Vision-and-Language Navigation

Vision-and-Language Navigation (VLN) tasks have gained prominence within artificial intelligence research due to their potential application in fields like home assistants. Many contemporary VLN approaches, while based on transformer architectures, have increasingly incorporated additional components such as external knowledge bases or map information to enhance performance. These additions, while boosting performance, also lead to larger models and increased computational costs. In this paper, to achieve both high performance and low computational costs, we propose a novel architecture with the COmbination of Selective MemOrization (COSMO). Specifically, COSMO integrates state-space modules and transformer modules, and incorporates two VLN-customized selective state space modules: the Round Selective Scan (RSS) and the Cross-modal Selective State Space Module (CS3). RSS facilitates comprehensive inter-modal interactions within a single scan, while the CS3 module adapts the selective state space module into a dual-stream architecture, thereby enhancing the acquisition of cross-modal interactions. Experimental validations on three mainstream VLN benchmarks, REVERIE, R2R, and R2R-CE, not only demonstrate competitive navigation performance of our model but also show a significant reduction in computational costs.

cs.CV↗

arXiv2024

Understanding the Natural Language of DNA using Encoder-Decoder Foundation Models with Byte-level Precision

This paper presents the Ensemble Nucleotide Byte-level Encoder-Decoder (ENBED) foundation model, analyzing DNA sequences at byte-level precision with an encoder-decoder Transformer architecture. ENBED uses a sub-quadratic implementation of attention to develop an efficient model capable of sequence-to-sequence transformations, generalizing previous genomic models with encoder-only or decoder-only architectures. We use Masked Language Modeling to pre-train the foundation model using reference genome sequences and apply it in the following downstream tasks: (1) identification of enhancers, promotors and splice sites, (2) recognition of sequences containing base call mismatches and insertion/deletion errors, an advantage over tokenization schemes involving multiple base pairs, which lose the ability to analyze with byte-level precision, (3) identification of biological function annotations of genomic sequences, and (4) generating mutations of the Influenza virus using the encoder-decoder architecture and validating them against real-world observations. In each of these tasks, we demonstrate significant improvement as compared to the existing state-of-the-art results.

cs.LG↗

arXiv2021

Geometry-Free View Synthesis: Transformers and no 3D Priors

Is a geometric model required to synthesize novel views from a single image? Being bound to local convolutions, CNNs need explicit 3D biases to model geometric transformations. In contrast, we demonstrate that a transformer-based model can synthesize entirely novel views without any hand-engineered 3D biases. This is achieved by (i) a global attention mechanism for implicitly learning long-range 3D correspondences between source and target views, and (ii) a probabilistic formulation necessary to capture the ambiguity inherent in predicting novel views from a single image, thereby overcoming the limitations of previous approaches that are restricted to relatively small viewpoint changes. We evaluate various ways to integrate 3D priors into a transformer architecture. However, our experiments show that no such geometric priors are required and that the transformer is capable of implicitly learning 3D relationships between images. Furthermore, this approach outperforms the state of the art in terms of visual quality while covering the full distribution of possible realizations. Code is available at https://git.io/JOnwn

cs.CV↗

arXiv2021

Probing Inter-modality: Visual Parsing with Self-Attention for Vision-Language Pre-training

Vision-Language Pre-training (VLP) aims to learn multi-modal representations from image-text pairs and serves for downstream vision-language tasks in a fine-tuning fashion. The dominant VLP models adopt a CNN-Transformer architecture, which embeds images with a CNN, and then aligns images and text with a Transformer. Visual relationship between visual contents plays an important role in image understanding and is the basic for inter-modal alignment learning. However, CNNs have limitations in visual relation learning due to local receptive field's weakness in modeling long-range dependencies. Thus the two objectives of learning visual relation and inter-modal alignment are encapsulated in the same Transformer network. Such design might restrict the inter-modal alignment learning in the Transformer by ignoring the specialized characteristic of each objective. To tackle this, we propose a fully Transformer visual embedding for VLP to better learn visual relation and further promote inter-modal alignment. Specifically, we propose a metric named Inter-Modality Flow (IMF) to measure the interaction between vision and language modalities (i.e., inter-modality). We also design a novel masking optimization mechanism named Masked Feature Regression (MFR) in Transformer to further promote the inter-modality learning. To the best of our knowledge, this is the first study to explore the benefit of Transformer for visual feature learning in VLP. We verify our method on a wide range of vision-language tasks, including Image-Text Retrieval, Visual Question Answering (VQA), Visual Entailment and Visual Reasoning. Our approach not only outperforms the state-of-the-art VLP performance, but also shows benefits on the IMF metric.

cs.CV↗

arXiv2022

TNT: Vision Transformer for Turbulence Simulations

Turbulence is notoriously difficult to model due to its multi-scale nature and sensitivity to small perturbations. Classical solvers of turbulence simulation generally operate on finer grids and are computationally inefficient. In this paper, we propose the Turbulence Neural Transformer (TNT), which is a learned simulator based on the transformer architecture, to predict turbulent dynamics on coarsened grids. TNT extends the positional embeddings of vanilla transformers to a spatiotemporal setting to learn the representation in the 3D time-series domain, and applies Temporal Mutual Self-Attention (TMSA), which captures adjacent dependencies, to extract deep and dynamic features. TNT is capable of generating comparatively long-range predictions stably and accurately, and we show that TNT outperforms the state-of-the-art U-net simulator on several metrics. We also test the model performance with different components removed and evaluate robustness to different initial conditions. Although more experiments are needed, we conclude that TNT has great potential to outperform existing solvers and generalize to additional simulation datasets.

physics.flu-dyn↗

arXiv2022

SkexGen: Autoregressive Generation of CAD Construction Sequences with Disentangled Codebooks

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.

cs.CV↗

arXiv2022

3D Vision with Transformers: A Survey

The success of the transformer architecture in natural language processing has recently triggered attention in the computer vision field. The transformer has been used as a replacement for the widely used convolution operators, due to its ability to learn long-range dependencies. This replacement was proven to be successful in numerous tasks, in which several state-of-the-art methods rely on transformers for better learning. In computer vision, the 3D field has also witnessed an increase in employing the transformer for 3D convolution neural networks and multi-layer perceptron networks. Although a number of surveys have focused on transformers in vision in general, 3D vision requires special attention due to the difference in data representation and processing when compared to 2D vision. In this work, we present a systematic and thorough review of more than 100 transformers methods for different 3D vision tasks, including classification, segmentation, detection, completion, pose estimation, and others. We discuss transformer design in 3D vision, which allows it to process data with various 3D representations. For each application, we highlight key properties and contributions of proposed transformer-based methods. To assess the competitiveness of these methods, we compare their performance to common non-transformer methods on 12 3D benchmarks. We conclude the survey by discussing different open directions and challenges for transformers in 3D vision. In addition to the presented papers, we aim to frequently update the latest relevant papers along with their corresponding implementations at: https://github.com/lahoud/3d-vision-transformers.

cs.CV↗

arXiv2025

A Survey on Long Text Modeling with Transformers

Modeling long texts has been an essential technique in the field of natural language processing (NLP). With the ever-growing number of long documents, it is important to develop effective modeling methods that can process and analyze such texts. However, long texts pose important research challenges for existing text models, with more complex semantics and special characteristics. In this paper, we provide an overview of the recent advances on long texts modeling based on Transformer models. Firstly, we introduce the formal definition of long text modeling. Then, as the core content, we discuss how to process long input to satisfy the length limitation and design improved Transformer architectures to effectively extend the maximum context length. Following this, we discuss how to adapt Transformer models to capture the special characteristics of long texts. Finally, we describe four typical applications involving long text modeling and conclude this paper with a discussion of future directions. Our survey intends to provide researchers with a synthesis and pointer to related work on long text modeling.

cs.CL↗

arXiv2023

QUIZ: An Arbitrary Volumetric Point Matching Method for Medical Image Registration

Rigid pre-registration involving local-global matching or other large deformation scenarios is crucial. Current popular methods rely on unsupervised learning based on grayscale similarity, but under circumstances where different poses lead to varying tissue structures, or where image quality is poor, these methods tend to exhibit instability and inaccuracies. In this study, we propose a novel method for medical image registration based on arbitrary voxel point of interest matching, called query point quizzer (QUIZ). QUIZ focuses on the correspondence between local-global matching points, specifically employing CNN for feature extraction and utilizing the Transformer architecture for global point matching queries, followed by applying average displacement for local image rigid transformation. We have validated this approach on a large deformation dataset of cervical cancer patients, with results indicating substantially smaller deviations compared to state-of-the-art methods. Remarkably, even for cross-modality subjects, it achieves results surpassing the current state-of-the-art.

cs.CV↗