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 685 records · Page 38Linked to original sources

arXiv2023

Full Point Encoding for Local Feature Aggregation in 3D Point Clouds

Point cloud processing methods exploit local point features and global context through aggregation which does not explicity model the internal correlations between local and global features. To address this problem, we propose full point encoding which is applicable to convolution and transformer architectures. Specifically, we propose Full Point Convolution (FPConv) and Full Point Transformer (FPTransformer) architectures. The key idea is to adaptively learn the weights from local and global geometric connections, where the connections are established through local and global correlation functions respectively. FPConv and FPTransformer simultaneously model the local and global geometric relationships as well as their internal correlations, demonstrating strong generalization ability and high performance. FPConv is incorporated in classical hierarchical network architectures to achieve local and global shape-aware learning. In FPTransformer, we introduce full point position encoding in self-attention, that hierarchically encodes each point position in the global and local receptive field. We also propose a shape aware downsampling block which takes into account the local shape and the global context. Experimental comparison to existing methods on benchmark datasets show the efficacy of FPConv and FPTransformer for semantic segmentation, object detection, classification, and normal estimation tasks. In particular, we achieve state-of-the-art semantic segmentation results of 76% mIoU on S3DIS 6-fold and 72.2% on S3DIS Area5.

cs.CV↗

arXiv2024

Advances in Transformers for Robotic Applications: A Review

The introduction of Transformers architecture has brought about significant breakthroughs in Deep Learning (DL), particularly within Natural Language Processing (NLP). Since their inception, Transformers have outperformed many traditional neural network architectures due to their "self-attention" mechanism and their scalability across various applications. In this paper, we cover the use of Transformers in Robotics. We go through recent advances and trends in Transformer architectures and examine their integration into robotic perception, planning, and control for autonomous systems. Furthermore, we review past work and recent research on use of Transformers in Robotics as pre-trained foundation models and integration of Transformers with Deep Reinforcement Learning (DRL) for autonomous systems. We discuss how different Transformer variants are being adapted in robotics for reliable planning and perception, increasing human-robot interaction, long-horizon decision-making, and generalization. Finally, we address limitations and challenges, offering insight and suggestions for future research directions.

cs.RO↗

arXiv2025

SegResMamba: An Efficient Architecture for 3D Medical Image Segmentation

The Transformer architecture has opened a new paradigm in the domain of deep learning with its ability to model long-range dependencies and capture global context and has outpaced the traditional Convolution Neural Networks (CNNs) in many aspects. However, applying Transformer models to 3D medical image datasets presents significant challenges due to their high training time, and memory requirements, which not only hinder scalability but also contribute to elevated CO$_2$ footprint. This has led to an exploration of alternative models that can maintain or even improve performance while being more efficient and environmentally sustainable. Recent advancements in Structured State Space Models (SSMs) effectively address some of the inherent limitations of Transformers, particularly their high memory and computational demands. Inspired by these advancements, we propose an efficient 3D segmentation model for medical imaging called SegResMamba, designed to reduce computation complexity, memory usage, training time, and environmental impact while maintaining high performance. Our model uses less than half the memory during training compared to other state-of-the-art (SOTA) architectures, achieving comparable performance with significantly reduced resource demands.

cs.CV↗

arXiv2025

JoFormer (Journey-based Transformer): Theory and Empirical Analysis on the Tiny Shakespeare Dataset

Transformers have demonstrated remarkable success in sequence modeling, yet effectively incorporating positional information remains a challenging and active area of research. In this paper, we introduce JoFormer, a journey-based Transformer architecture grounded in a recently proposed non-commutative algebra for composing transformations across positions. JoFormer represents relative positions through learnable directional transforms that are sequentially composed along the input, thereby extending and generalizing existing approaches based on relative position representations. We derive the JoFormer attention mechanism from first principles and show that it subsumes standard methods such as rotary transformations as special cases. To evaluate its effectiveness, we compare JoFormer to the RoFormer baseline on the Tiny Shakespeare character-level language modeling task. Our results demonstrate that JoFormer consistently achieves lower perplexity and faster convergence, highlighting the advantages of its more expressive, journey-based treatment of position. Notably, the per-token JoFormer is still a primitive, conceptual variant with layer-independent angles, yet it already demonstrates strong performance-underscoring its promise as a proof of concept for more expressive architectures. We conclude by discussing how JoFormer offers a principled approach to integrating positional structure into Transformer architectures. The code used in this work is available at https://github.com/mahesh-godavarti/joformer.

cs.LG↗

arXiv2025

Leveraging Decoder Architectures for Learned Sparse Retrieval

Learned Sparse Retrieval (LSR) has traditionally focused on small-scale encoder-only transformer architectures. With the advent of large-scale pre-trained language models, their capability to generate sparse representations for retrieval tasks across different transformer-based architectures, including encoder-only, decoder-only, and encoder-decoder models, remains largely unexplored. This study investigates the effectiveness of LSR across these architectures, exploring various sparse representation heads and model scales. Our results highlight the limitations of using large language models to create effective sparse representations in zero-shot settings, identifying challenges such as inappropriate term expansions and reduced performance due to the lack of expansion. We find that the encoder-decoder architecture with multi-tokens decoding approach achieves the best performance among the three backbones. While the decoder-only model performs worse than the encoder-only model, it demonstrates the potential to outperform when scaled to a high number of parameters.

cs.IR↗

arXiv2026

Image-to-Brain Signal Generation for Visual Prosthesis with CLIP Guided Multimodal Diffusion Models

Visual prostheses hold great promise for restoring vision in blind individuals. While researchers have successfully utilized M/EEG signals to evoke visual perceptions during the brain decoding stage of visual prostheses, the complementary process of converting images into M/EEG signals in the brain encoding stage remains largely unexplored, hindering the formation of a complete functional pipeline. In this work, we present a novel image-to-brain signal framework that generates M/EEG from images by leveraging the diffusion transformer architecture enhanced with cross-attention mechanisms. Specifically, we employ a diffusion transformer (DiT) architecture based on denoising diffusion implicit models (DDIM) to achieve brain signal generation. To realize the goal of image-to-brain signal conversion, we use cross-attention mechanisms to align brain signal embeddings with CLIP image embeddings. Moreover, we leverage large language models (LLMs) to generate image captions, and concatenate the resulting CLIP text embeddings with CLIP image embeddings to form unified embeddings for cross-attention alignment, enabling our model to capture core semantic information. Moreover, to capture core semantic information, we use large language models (LLMs) to generate descriptive and semantically accurate captions for images. Furthermore, we introduce a learnable spatio-temporal position encoding that combines brain region embeddings with temporal embeddings to capture both spatial and temporal characteristics of brain signals. We evaluate the framework on two multimodal benchmark datasets (THINGS-EEG2 and THINGS-MEG) and demonstrate that it generates biologically plausible brain signals.

cs.CV↗

arXiv2026

Pattern Formation in Transformers

What are the inductive biases of a Transformer architecture? Existing theory on how the forward pass shapes representations either considers whether Transformers escape from rank collapse or demonstrates that self-attention drives tokens toward cluster patterns. The latter view arises from an elegant dynamical systems perspective, but relies on simplified architectural assumptions, and does not explain the rich structures observed in practice. This leaves a major open question: when a full Transformer escapes rank collapse, how does it structure token representations? Using pattern-formation theory, we show that the dynamical view of Transformers can account for Positional Encoding, Multi-Head Attention, and Output-Value geometry. We demonstrate that a full Transformer architecture imposes an inductive prior by selectively amplifying a rich set of previously unreported patterns, including traveling or rotating waves among others. We characterize the role of each architectural component in controlling which pattern is amplified, which ones stabilize, compete, or coexist. Finally, we show that these structures can act as a controllable dynamical prior that facilitates learning. By choosing both task-aligned positional encoding and weight initialization, we demonstrate improved data efficiency and accelerated optimization on controlled sequence tasks and with ConViT on CIFAR-10.

cs.LG↗

arXiv2022

TraHGR: Transformer for Hand Gesture Recognition via ElectroMyography

Deep learning-based Hand Gesture Recognition (HGR) via surface Electromyogram (sEMG) signals has recently shown significant potential for development of advanced myoelectric-controlled prosthesis. Existing deep learning approaches, typically, include only one model as such can hardly maintain acceptable generalization performance in changing scenarios. In this paper, we aim to address this challenge by capitalizing on the recent advances of hybrid models and transformers. In other words, we propose a hybrid framework based on the transformer architecture, which is a relatively new and revolutionizing deep learning model. The proposed hybrid architecture, referred to as the Transformer for Hand Gesture Recognition (TraHGR), consists of two parallel paths followed by a linear layer that acts as a fusion center to integrate the advantage of each module and provide robustness over different scenarios. We evaluated the proposed architecture TraHGR based on the commonly used second Ninapro dataset, referred to as the DB2. The sEMG signals in the DB2 dataset are measured in the real-life conditions from 40 healthy users, each performing 49 gestures. We have conducted extensive set of experiments to test and validate the proposed TraHGR architecture, and have compared its achievable accuracy with more than five recently proposed HGR classification algorithms over the same dataset. We have also compared the results of the proposed TraHGR architecture with each individual path and demonstrated the distinguishing power of the proposed hybrid architecture. The recognition accuracies of the proposed TraHGR architecture are 86.18%, 88.91%, 81.44%, and 93.84%, which are 2.48%, 5.12%, 8.82%, and 4.30% higher than the state-ofthe-art performance for DB2 (49 gestures), DB2-B (17 gestures), DB2-C (23 gestures), and DB2-D (9 gestures), respectively.

eess.SP↗

arXiv2026

Dual Filter: A Transformer-like Inference Architecture for Hidden Markov Models

This paper presents a mathematical framework for causal nonlinear prediction in settings where observations are generated from an underlying hidden Markov model (HMM). Both the problem formulation and the proposed solution are motivated by the decoder-only transformer architecture, in which a finite sequence of observations (tokens) is mapped to the conditional probability of the next token. Our objective is not to construct a mathematical model of a transformer. Rather, our interest lies in deriving, from first principles, transformer-like architectures that solve the prediction problem for which the transformer is designed. The proposed framework is based on an original optimal control approach, where the prediction objective (MMSE) is reformulated as an optimal control problem. An analysis of the optimal control problem is presented leading to a fixed-point equation on the space of probability measures. To solve the fixed-point equation, we introduce the dual filter, an iterative algorithm that closely parallels the architecture of decoder-only transformers. These parallels are discussed in detail along with the relationship to prior work on mathematical modeling of transformers as transport on the space of probability measures. Numerical experiments are provided to illustrate the performance of the algorithm using parameter values typical of research-scale transformer models.

cs.LG↗

arXiv2025

FullTransNet: Full Transformer with Local-Global Attention for Video Summarization

Video summarization aims to generate a compact, informative, and representative synopsis of raw videos, which is crucial for browsing, analyzing, and understanding video content. Dominant approaches in video summarization primarily rely on recurrent or convolutional neural networks, and more recently on encoder-only transformer architectures. However, these methods typically suffer from several limitations in parallelism, modeling long-range dependencies, and providing explicit generative capabilities. To address these issues, we propose a transformer-like architecture named FullTransNet with two-fold ideas. First, it uses a full transformer with an encoder-decoder structure as an alternative architecture for video summarization. As the full transformer is specifically designed for sequence transduction tasks, its direct application to video summarization is both intuitive and effective. Second, it replaces the standard full attention mechanism with a combination of local and global sparse attention, enabling the model to capture long-range dependencies while significantly reducing computational costs. This local-global sparse attention is applied exclusively at the encoder side, where the majority of computations occur, further enhancing efficiency. Extensive experiments on two widely used benchmark datasets, SumMe and TVSum, demonstrate that our model achieves F-scores of 54.4% and 63.9%, respectively, while maintaining relatively low computational and memory requirements. These results surpass the second-best performing methods by 0.1% and 0.3%, respectively, verifying the effectiveness and efficiency of FullTransNet.

cs.CV↗

arXiv2021

Condenser: a Pre-training Architecture for Dense Retrieval

Pre-trained Transformer language models (LM) have become go-to text representation encoders. Prior research fine-tunes deep LMs to encode text sequences such as sentences and passages into single dense vector representations for efficient text comparison and retrieval. However, dense encoders require a lot of data and sophisticated techniques to effectively train and suffer in low data situations. This paper finds a key reason is that standard LMs' internal attention structure is not ready-to-use for dense encoders, which needs to aggregate text information into the dense representation. We propose to pre-train towards dense encoder with a novel Transformer architecture, Condenser, where LM prediction CONditions on DENSE Representation. Our experiments show Condenser improves over standard LM by large margins on various text retrieval and similarity tasks.

cs.CL↗

arXiv2023

TransCODE: Co-design of Transformers and Accelerators for Efficient Training and Inference

Automated co-design of machine learning models and evaluation hardware is critical for efficiently deploying such models at scale. Despite the state-of-the-art performance of transformer models, they are not yet ready for execution on resource-constrained hardware platforms. High memory requirements and low parallelizability of the transformer architecture exacerbate this problem. Recently-proposed accelerators attempt to optimize the throughput and energy consumption of transformer models. However, such works are either limited to a one-sided search of the model architecture or a restricted set of off-the-shelf devices. Furthermore, previous works only accelerate model inference and not training, which incurs substantially higher memory and compute resources, making the problem even more challenging. To address these limitations, this work proposes a dynamic training framework, called DynaProp, that speeds up the training process and reduces memory consumption. DynaProp is a low-overhead pruning method that prunes activations and gradients at runtime. To effectively execute this method on hardware for a diverse set of transformer architectures, we propose ELECTOR, a framework that simulates transformer inference and training on a design space of accelerators. We use this simulator in conjunction with the proposed co-design technique, called TransCODE, to obtain the best-performing models with high accuracy on the given task and minimize latency, energy consumption, and chip area. The obtained transformer-accelerator pair achieves 0.3% higher accuracy than the state-of-the-art pair while incurring 5.2$\times$ lower latency and 3.0$\times$ lower energy consumption.

cs.LG↗

arXiv2021

Progressive Transformer-Based Generation of Radiology Reports

Inspired by Curriculum Learning, we propose a consecutive (i.e., image-to-text-to-text) generation framework where we divide the problem of radiology report generation into two steps. Contrary to generating the full radiology report from the image at once, the model generates global concepts from the image in the first step and then reforms them into finer and coherent texts using a transformer architecture. We follow the transformer-based sequence-to-sequence paradigm at each step. We improve upon the state-of-the-art on two benchmark datasets.

cs.CL↗

arXiv2021

Learning Attributed Graph Representations with Communicative Message Passing Transformer

Constructing appropriate representations of molecules lies at the core of numerous tasks such as material science, chemistry and drug designs. Recent researches abstract molecules as attributed graphs and employ graph neural networks (GNN) for molecular representation learning, which have made remarkable achievements in molecular graph modeling. Albeit powerful, current models either are based on local aggregation operations and thus miss higher-order graph properties or focus on only node information without fully using the edge information. For this sake, we propose a Communicative Message Passing Transformer (CoMPT) neural network to improve the molecular graph representation by reinforcing message interactions between nodes and edges based on the Transformer architecture. Unlike the previous transformer-style GNNs that treat molecules as fully connected graphs, we introduce a message diffusion mechanism to leverage the graph connectivity inductive bias and reduce the message enrichment explosion. Extensive experiments demonstrated that the proposed model obtained superior performances (around 4$\%$ on average) against state-of-the-art baselines on seven chemical property datasets (graph-level tasks) and two chemical shift datasets (node-level tasks). Further visualization studies also indicated a better representation capacity achieved by our model.

cs.LG↗

arXiv2022

Unleashing the Power of Transformer for Graphs

Despite recent successes in natural language processing and computer vision, Transformer suffers from the scalability problem when dealing with graphs. The computational complexity is unacceptable for large-scale graphs, e.g., knowledge graphs. One solution is to consider only the near neighbors, which, however, will lose the key merit of Transformer to attend to the elements at any distance. In this paper, we propose a new Transformer architecture, named dual-encoding Transformer (DET). DET has a structural encoder to aggregate information from connected neighbors and a semantic encoder to focus on semantically useful distant nodes. In comparison with resorting to multi-hop neighbors, DET seeks the desired distant neighbors via self-supervised training. We further find these two encoders can be incorporated to boost each others' performance. Our experiments demonstrate DET has achieved superior performance compared to the respective state-of-the-art methods in dealing with molecules, networks and knowledge graphs with various sizes.

cs.LG↗

arXiv2023

The Parallelism Tradeoff: Limitations of Log-Precision Transformers

Despite their omnipresence in modern NLP, characterizing the computational power of transformer neural nets remains an interesting open question. We prove that transformers whose arithmetic precision is logarithmic in the number of input tokens (and whose feedforward nets are computable using space linear in their input) can be simulated by constant-depth logspace-uniform threshold circuits. This provides insight on the power of transformers using known results in complexity theory. For example, if $\mathsf L \neq \mathsf P$ (i.e., not all poly-time problems can be solved using logarithmic space), then transformers cannot even accurately solve linear equalities or check membership in an arbitrary context-free grammar with empty productions. Our result intuitively emerges from the transformer architecture's high parallelizability. We thus speculatively introduce the idea of a fundamental parallelism tradeoff: any model architecture as parallelizable as the transformer will obey limitations similar to it. Since parallelism is key to training models at massive scale, this suggests a potential inherent weakness of the scaling paradigm.

cs.CC↗

arXiv2022

Transformers for Object Detection in Large Point Clouds

We present TransLPC, a novel detection model for large point clouds that is based on a transformer architecture. While object detection with transformers has been an active field of research, it has proved difficult to apply such models to point clouds that span a large area, e.g. those that are common in autonomous driving, with lidar or radar data. TransLPC is able to remedy these issues: The structure of the transformer model is modified to allow for larger input sequence lengths, which are sufficient for large point clouds. Besides this, we propose a novel query refinement technique to improve detection accuracy, while retaining a memory-friendly number of transformer decoder queries. The queries are repositioned between layers, moving them closer to the bounding box they are estimating, in an efficient manner. This simple technique has a significant effect on detection accuracy, which is evaluated on the challenging nuScenes dataset on real-world lidar data. Besides this, the proposed method is compatible with existing transformer-based solutions that require object detection, e.g. for joint multi-object tracking and detection, and enables them to be used in conjunction with large point clouds.

cs.CV↗

arXiv2024

Q-HyViT: Post-Training Quantization of Hybrid Vision Transformers with Bridge Block Reconstruction for IoT Systems

Recently, vision transformers (ViTs) have superseded convolutional neural networks in numerous applications, including classification, detection, and segmentation. However, the high computational requirements of ViTs hinder their widespread implementation. To address this issue, researchers have proposed efficient hybrid transformer architectures that combine convolutional and transformer layers with optimized attention computation of linear complexity. Additionally, post-training quantization has been proposed as a means of mitigating computational demands. For mobile devices, achieving optimal acceleration for ViTs necessitates the strategic integration of quantization techniques and efficient hybrid transformer structures. However, no prior investigation has applied quantization to efficient hybrid transformers. In this paper, we discover that applying existing post-training quantization (PTQ) methods for ViTs to efficient hybrid transformers leads to a drastic accuracy drop, attributed to the four following challenges: (i) highly dynamic ranges, (ii) zero-point overflow, (iii) diverse normalization, and (iv) limited model parameters ($<$5M). To overcome these challenges, we propose a new post-training quantization method, which is the first to quantize efficient hybrid ViTs (MobileViTv1, MobileViTv2, Mobile-Former, EfficientFormerV1, EfficientFormerV2). We achieve a significant improvement of 17.73% for 8-bit and 29.75% for 6-bit on average, respectively, compared with existing PTQ methods (EasyQuant, FQ-ViT, PTQ4ViT, and RepQ-ViT)}. We plan to release our code at https://gitlab.com/ones-ai/q-hyvit.

cs.CV↗