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 559 records · Page 31Linked to original sources

arXiv2024

Towards Multi-modal Transformers in Federated Learning

Multi-modal transformers mark significant progress in different domains, but siloed high-quality data hinders their further improvement. To remedy this, federated learning (FL) has emerged as a promising privacy-preserving paradigm for training models without direct access to the raw data held by different clients. Despite its potential, a considerable research direction regarding the unpaired uni-modal clients and the transformer architecture in FL remains unexplored. To fill this gap, this paper explores a transfer multi-modal federated learning (MFL) scenario within the vision-language domain, where clients possess data of various modalities distributed across different datasets. We systematically evaluate the performance of existing methods when a transformer architecture is utilized and introduce a novel framework called Federated modality complementary and collaboration (FedCola) by addressing the in-modality and cross-modality gaps among clients. Through extensive experiments across various FL settings, FedCola demonstrates superior performance over previous approaches, offering new perspectives on future federated training of multi-modal transformers.

cs.CV↗

arXiv2026

A statistical perspective on transformers for small longitudinal cohort data

Modeling of longitudinal cohort data typically involves complex temporal dependencies between multiple variables. There, the transformer architecture, which has been highly successful in language and vision applications, allows us to account for the fact that the most recently observed time points in an individual's history may not always be the most important for the immediate future. This is achieved by assigning attention weights to observations of an individual based on a transformation of their values. One reason why these ideas have not yet been fully leveraged for longitudinal cohort data is that typically, large datasets are required. Therefore, we present a simplified transformer architecture that retains the core attention mechanism while reducing the number of parameters to be estimated, to be more suitable for small datasets with few time points. Guided by a statistical perspective on transformers, we use an autoregressive model as a starting point and incorporate attention as a kernel-based operation with temporal decay, where aggregation of multiple transformer heads, i.e. different candidate weighting schemes, is expressed as accumulating evidence on different types of underlying characteristics of individuals. This also enables a permutation-based statistical testing procedure for identifying contextual patterns. In a simulation study, the approach is shown to recover contextual dependencies even with a small number of individuals and time points. In an application to data from a resilience study, we identify temporal patterns in the dynamics of stress and mental health. This indicates that properly adapted transformers can not only achieve competitive predictive performance, but also uncover complex context dependencies in small data settings.

stat.ME↗

arXiv2023

BEL: A Bag Embedding Loss for Transformer enhances Multiple Instance Whole Slide Image Classification

Multiple Instance Learning (MIL) has become the predominant approach for classification tasks on gigapixel histopathology whole slide images (WSIs). Within the MIL framework, single WSIs (bags) are decomposed into patches (instances), with only WSI-level annotation available. Recent MIL approaches produce highly informative bag level representations by utilizing the transformer architecture's ability to model the dependencies between instances. However, when applied to high magnification datasets, problems emerge due to the large number of instances and the weak supervisory learning signal. To address this problem, we propose to additionally train transformers with a novel Bag Embedding Loss (BEL). BEL forces the model to learn a discriminative bag-level representation by minimizing the distance between bag embeddings of the same class and maximizing the distance between different classes. We evaluate BEL with the Transformer architecture TransMIL on two publicly available histopathology datasets, BRACS and CAMELYON17. We show that with BEL, TransMIL outperforms the baseline models on both datasets, thus contributing to the clinically highly relevant AI-based tumor classification of histological patient material.

cs.CV↗

arXiv2023

Transformers as Statisticians: Provable In-Context Learning with In-Context Algorithm Selection

Neural sequence models based on the transformer architecture have demonstrated remarkable \emph{in-context learning} (ICL) abilities, where they can perform new tasks when prompted with training and test examples, without any parameter update to the model. This work first provides a comprehensive statistical theory for transformers to perform ICL. Concretely, we show that transformers can implement a broad class of standard machine learning algorithms in context, such as least squares, ridge regression, Lasso, learning generalized linear models, and gradient descent on two-layer neural networks, with near-optimal predictive power on various in-context data distributions. Using an efficient implementation of in-context gradient descent as the underlying mechanism, our transformer constructions admit mild size bounds, and can be learned with polynomially many pretraining sequences. Building on these ``base'' ICL algorithms, intriguingly, we show that transformers can implement more complex ICL procedures involving \emph{in-context algorithm selection}, akin to what a statistician can do in real life -- A \emph{single} transformer can adaptively select different base ICL algorithms -- or even perform qualitatively different tasks -- on different input sequences, without any explicit prompting of the right algorithm or task. We both establish this in theory by explicit constructions, and also observe this phenomenon experimentally. In theory, we construct two general mechanisms for algorithm selection with concrete examples: pre-ICL testing, and post-ICL validation. As an example, we use the post-ICL validation mechanism to construct a transformer that can perform nearly Bayes-optimal ICL on a challenging task -- noisy linear models with mixed noise levels. Experimentally, we demonstrate the strong in-context algorithm selection capabilities of standard transformer architectures.

cs.LG↗

arXiv2025

Brain-inspired Action Generation with Spiking Transformer Diffusion Policy Model

Spiking Neural Networks (SNNs) has the ability to extract spatio-temporal features due to their spiking sequence. While previous research has primarily foucus on the classification of image and reinforcement learning. In our paper, we put forward novel diffusion policy model based on Spiking Transformer Neural Networks and Denoising Diffusion Probabilistic Model (DDPM): Spiking Transformer Modulate Diffusion Policy Model (STMDP), a new brain-inspired model for generating robot action trajectories. In order to improve the performance of this model, we develop a novel decoder module: Spiking Modulate De coder (SMD), which replaces the traditional Decoder module within the Transformer architecture. Additionally, we explored the substitution of DDPM with Denoising Diffusion Implicit Models (DDIM) in our frame work. We conducted experiments across four robotic manipulation tasks and performed ablation studies on the modulate block. Our model consistently outperforms existing Transformer-based diffusion policy method. Especially in Can task, we achieved an improvement of 8%. The proposed STMDP method integrates SNNs, dffusion model and Transformer architecture, which offers new perspectives and promising directions for exploration in brain-inspired robotics.

cs.RO↗

arXiv2025

VESTA: A Versatile SNN-Based Transformer Accelerator with Unified PEs for Multiple Computational Layers

Spiking Neural Networks (SNNs) and transformers represent two powerful paradigms in neural computation, known for their low power consumption and ability to capture feature dependencies, respectively. However, transformer architectures typically involve multiple types of computational layers, including linear layers for MLP modules and classification heads, convolution layers for tokenizers, and dot product computations for self-attention mechanisms. These diverse operations pose significant challenges for hardware accelerator design, and to our knowledge, there is not yet a hardware solution that leverages spike-form data from SNNs for transformer architectures. In this paper, we introduce VESTA, a novel hardware design that synergizes these technologies, presenting unified Processing Elements (PEs) capable of efficiently performing all three types of computations crucial to transformer structures. VESTA uniquely benefits from the spike-form outputs of the Spike Neuron Layers \cite{zhou2024spikformer}, simplifying multiplication operations by reducing them from handling two 8-bit integers to handling one 8-bit integer and a binary spike. This reduction enables the use of multiplexers in the PE module, significantly enhancing computational efficiency while maintaining the low-power advantage of SNNs. Experimental results show that the core area of VESTA is $0.844 mm^2$. It operates at 500MHz and is capable of real-time image classification at 30 fps.

cs.AR↗

arXiv2022

Transformers are Meta-Reinforcement Learners

The transformer architecture and variants presented remarkable success across many machine learning tasks in recent years. This success is intrinsically related to the capability of handling long sequences and the presence of context-dependent weights from the attention mechanism. We argue that these capabilities suit the central role of a Meta-Reinforcement Learning algorithm. Indeed, a meta-RL agent needs to infer the task from a sequence of trajectories. Furthermore, it requires a fast adaptation strategy to adapt its policy for a new task -- which can be achieved using the self-attention mechanism. In this work, we present TrMRL (Transformers for Meta-Reinforcement Learning), a meta-RL agent that mimics the memory reinstatement mechanism using the transformer architecture. It associates the recent past of working memories to build an episodic memory recursively through the transformer layers. We show that the self-attention computes a consensus representation that minimizes the Bayes Risk at each layer and provides meaningful features to compute the best actions. We conducted experiments in high-dimensional continuous control environments for locomotion and dexterous manipulation. Results show that TrMRL presents comparable or superior asymptotic performance, sample efficiency, and out-of-distribution generalization compared to the baselines in these environments.

cs.LG↗

arXiv2025

Context-Aware Rule Mining Using a Dynamic Transformer-Based Framework

This study proposes a dynamic rule data mining algorithm based on an improved Transformer architecture, aiming to improve the accuracy and efficiency of rule mining in a dynamic data environment. With the increase in data volume and complexity, traditional data mining methods are difficult to cope with dynamic data with strong temporal and variable characteristics, so new algorithms are needed to capture the temporal regularity in the data. By improving the Transformer architecture, and introducing a dynamic weight adjustment mechanism and a temporal dependency module, we enable the model to adapt to data changes and mine more accurate rules. Experimental results show that compared with traditional rule mining algorithms, the improved Transformer model has achieved significant improvements in rule mining accuracy, coverage, and stability. The contribution of each module in the algorithm performance is further verified by ablation experiments, proving the importance of temporal dependency and dynamic weight adjustment mechanisms in improving the model effect. In addition, although the improved model has certain challenges in computational efficiency, its advantages in accuracy and coverage enable it to perform well in processing complex dynamic data. Future research will focus on optimizing computational efficiency and combining more deep learning technologies to expand the application scope of the algorithm, especially in practical applications in the fields of finance, medical care, and intelligent recommendation.

cs.LG↗

arXiv2025

IncepFormerNet: A multi-scale multi-head attention network for SSVEP classification

In recent years, deep learning (DL) models have shown outstanding performance in EEG classification tasks, particularly in Steady-State Visually Evoked Potential(SSVEP)-based Brain-Computer-Interfaces(BCI)systems. DL methods have been successfully applied to SSVEP-BCI. This study proposes a new model called IncepFormerNet, which is a hybrid of the Inception and Transformer architectures. IncepFormerNet adeptly extracts multi-scale temporal information from time series data using parallel convolution kernels of varying sizes, accurately capturing the subtle variations and critical features within SSVEP signals.Furthermore, the model integrates the multi-head attention mechanism from the Transformer architecture, which not only provides insights into global dependencies but also significantly enhances the understanding and representation of complex patterns.Additionally, it takes advantage of filter bank techniques to extract features based on the spectral characteristics of SSVEP data. To validate the effectiveness of the proposed model, we conducted experiments on two public datasets, . The experimental results show that IncepFormerNet achieves an accuracy of 87.41 on Dataset 1 and 71.97 on Dataset 2 using a 1.0-second time window. To further verify the superiority of the proposed model, we compared it with other deep learning models, and the results indicate that our method achieves significantly higher accuracy than the others.The source codes in this work are available at: https://github.com/CECNL/SSVEP-DAN.

eess.SP↗

arXiv2026

SpanNorm: Reconciling Training Stability and Performance in Deep Transformers

The success of Large Language Models (LLMs) hinges on the stable training of deep Transformer architectures. A critical design choice is the placement of normalization layers, leading to a fundamental trade-off: the ``PreNorm'' architecture ensures training stability at the cost of potential performance degradation in deep models, while the ``PostNorm'' architecture offers strong performance but suffers from severe training instability. In this work, we propose SpanNorm, a novel technique designed to resolve this dilemma by integrating the strengths of both paradigms. Structurally, SpanNorm establishes a clean residual connection that spans the entire transformer block to stabilize signal propagation, while employing a PostNorm-style computation that normalizes the aggregated output to enhance model performance. We provide a theoretical analysis demonstrating that SpanNorm, combined with a principled scaling strategy, maintains bounded signal variance throughout the network, preventing the gradient issues that plague PostNorm models, and also alleviating the representation collapse of PreNorm. Empirically, SpanNorm consistently outperforms standard normalization schemes in both dense and Mixture-of-Experts (MoE) scenarios, paving the way for more powerful and stable Transformer architectures.

cs.CL↗

arXiv2024

RATSF: Empowering Customer Service Volume Management through Retrieval-Augmented Time-Series Forecasting

An efficient customer service management system hinges on precise forecasting of service volume. In this scenario, where data non-stationarity is pronounced, successful forecasting heavily relies on identifying and leveraging similar historical data rather than merely summarizing periodic patterns. Existing models based on RNN or Transformer architectures may struggle with this flexible and effective utilization. To tackle this challenge, we initially developed the Time Series Knowledge Base (TSKB) with an advanced indexing system for efficient historical data retrieval. We also developed the Retrieval Augmented Cross-Attention (RACA) module, a variant of the cross-attention mechanism within Transformer's decoder layers, designed to be seamlessly integrated into the vanilla Transformer architecture to assimilate key historical data segments. The synergy between TSKB and RACA forms the backbone of our Retrieval-Augmented Time Series Forecasting (RATSF) framework. Based on the above two components, RATSF not only significantly enhances performance in the context of Fliggy hotel service volume forecasting but also adapts flexibly to various scenarios and integrates with a multitude of Transformer variants for time-series forecasting. Extensive experimentation has validated the effectiveness and generalizability of this system design across multiple diverse contexts.

cs.LG↗

arXiv2026

EmbBERT: Attention Under 2 MB Memory

Transformer architectures based on the attention mechanism have revolutionized natural language processing (NLP), driving major breakthroughs across virtually every NLP task. However, their substantial memory and computational requirements still hinder deployment on ultra-constrained devices such as wearables and Internet-of-Things (IoT) units, where available memory is limited to just a few megabytes. To address this challenge, we introduce EmbBERT, a tiny language model (TLM) architecturally designed for extreme efficiency. The model integrates a compact embedding layer, streamlined feed-forward blocks, and an efficient attention mechanism that together enable optimal performance under strict memory budgets. Through this redesign for the extreme edge, we demonstrate that highly simplified transformer architectures remain remarkably effective under tight resource constraints. EmbBERT requires only 2 MB of total memory, and achieves accuracy performance comparable to the ones of state-of-the-art (SotA) models that require a $\mathbf{10\times}$ memory budget. Extensive experiments on the curated TinyNLP benchmark and the GLUE suite confirm that EmbBERT achieves competitive accuracy, comparable to that of larger SotA models, and consistently outperforms downsized versions of BERT and MAMBA of similar size. Furthermore, we demonstrate the model resilience to 8-bit quantization, which further reduces memory usage to just 781 kB , and the scalability of the EmbBERT architecture across the sub-megabyte to tens-of-megabytes range. Finally, we perform an ablation study demonstrating the positive contributions of all components and the pre-training procedure. All code, scripts, and checkpoints are publicly released to ensure reproducibility: https://github.com/RiccardoBravin/tiny-LLM.

cs.CL↗

arXiv2024

Hybrid Quantum Vision Transformers for Event Classification in High Energy Physics

Models based on vision transformer architectures are considered state-of-the-art when it comes to image classification tasks. However, they require extensive computational resources both for training and deployment. The problem is exacerbated as the amount and complexity of the data increases. Quantum-based vision transformer models could potentially alleviate this issue by reducing the training and operating time while maintaining the same predictive power. Although current quantum computers are not yet able to perform high-dimensional tasks yet, they do offer one of the most efficient solutions for the future. In this work, we construct several variations of a quantum hybrid vision transformer for a classification problem in high energy physics (distinguishing photons and electrons in the electromagnetic calorimeter). We test them against classical vision transformer architectures. Our findings indicate that the hybrid models can achieve comparable performance to their classical analogues with a similar number of parameters.

quant-ph↗

arXiv2024

Attending to Graph Transformers

Recently, transformer architectures for graphs emerged as an alternative to established techniques for machine learning with graphs, such as (message-passing) graph neural networks. So far, they have shown promising empirical results, e.g., on molecular prediction datasets, often attributed to their ability to circumvent graph neural networks' shortcomings, such as over-smoothing and over-squashing. Here, we derive a taxonomy of graph transformer architectures, bringing some order to this emerging field. We overview their theoretical properties, survey structural and positional encodings, and discuss extensions for important graph classes, e.g., 3D molecular graphs. Empirically, we probe how well graph transformers can recover various graph properties, how well they can deal with heterophilic graphs, and to what extent they prevent over-squashing. Further, we outline open challenges and research direction to stimulate future work. Our code is available at https://github.com/luis-mueller/probing-graph-transformers.

cs.LG↗

arXiv2021

Dynamic Transformer for Efficient Machine Translation on Embedded Devices

The Transformer architecture is widely used for machine translation tasks. However, its resource-intensive nature makes it challenging to implement on constrained embedded devices, particularly where available hardware resources can vary at run-time. We propose a dynamic machine translation model that scales the Transformer architecture based on the available resources at any particular time. The proposed approach, 'Dynamic-HAT', uses a HAT SuperTransformer as the backbone to search for SubTransformers with different accuracy-latency trade-offs at design time. The optimal SubTransformers are sampled from the SuperTransformer at run-time, depending on latency constraints. The Dynamic-HAT is tested on the Jetson Nano and the approach uses inherited SubTransformers sampled directly from the SuperTransformer with a switching time of <1s. Using inherited SubTransformers results in a BLEU score loss of <1.5% because the SubTransformer configuration is not retrained from scratch after sampling. However, to recover this loss in performance, the dimensions of the design space can be reduced to tailor it to a family of target hardware. The new reduced design space results in a BLEU score increase of approximately 1% for sub-optimal models from the original design space, with a wide range for performance scaling between 0.356s - 1.526s for the GPU and 2.9s - 7.31s for the CPU.

cs.CL↗

arXiv2024

IPT-V2: Efficient Image Processing Transformer using Hierarchical Attentions

Recent advances have demonstrated the powerful capability of transformer architecture in image restoration. However, our analysis indicates that existing transformerbased methods can not establish both exact global and local dependencies simultaneously, which are much critical to restore the details and missing content of degraded images. To this end, we present an efficient image processing transformer architecture with hierarchical attentions, called IPTV2, adopting a focal context self-attention (FCSA) and a global grid self-attention (GGSA) to obtain adequate token interactions in local and global receptive fields. Specifically, FCSA applies the shifted window mechanism into the channel self-attention, helps capture the local context and mutual interaction across channels. And GGSA constructs long-range dependencies in the cross-window grid, aggregates global information in spatial dimension. Moreover, we introduce structural re-parameterization technique to feed-forward network to further improve the model capability. Extensive experiments demonstrate that our proposed IPT-V2 achieves state-of-the-art results on various image processing tasks, covering denoising, deblurring, deraining and obtains much better trade-off for performance and computational complexity than previous methods. Besides, we extend our method to image generation as latent diffusion backbone, and significantly outperforms DiTs.

cs.CV↗

arXiv2023

IDPL-PFOD2: A New Large-Scale Dataset for Printed Farsi Optical Character Recognition

Optical Character Recognition is a technique that converts document images into searchable and editable text, making it a valuable tool for processing scanned documents. While the Farsi language stands as a prominent and official language in Asia, efforts to develop efficient methods for recognizing Farsi printed text have been relatively limited. This is primarily attributed to the languages distinctive features, such as cursive form, the resemblance between certain alphabet characters, and the presence of numerous diacritics and dot placement. On the other hand, given the substantial training sample requirements of deep-based architectures for effective performance, the development of such datasets holds paramount significance. In light of these concerns, this paper aims to present a novel large-scale dataset, IDPL-PFOD2, tailored for Farsi printed text recognition. The dataset comprises 2003541 images featuring a wide variety of fonts, styles, and sizes. This dataset is an extension of the previously introduced IDPL-PFOD dataset, offering a substantial increase in both volume and diversity. Furthermore, the datasets effectiveness is assessed through the utilization of both CRNN-based and Vision Transformer architectures. The CRNN-based model achieves a baseline accuracy rate of 78.49% and a normalized edit distance of 97.72%, while the Vision Transformer architecture attains an accuracy of 81.32% and a normalized edit distance of 98.74%.

cs.CV↗

arXiv2025

GraphXForm: Graph transformer for computer-aided molecular design

Generative deep learning has become pivotal in molecular design for drug discovery, materials science, and chemical engineering. A widely used paradigm is to pretrain neural networks on string representations of molecules and fine-tune them using reinforcement learning on specific objectives. However, string-based models face challenges in ensuring chemical validity and enforcing structural constraints like the presence of specific substructures. We propose to instead combine graph-based molecular representations, which can naturally ensure chemical validity, with transformer architectures, which are highly expressive and capable of modeling long-range dependencies between atoms. Our approach iteratively modifies a molecular graph by adding atoms and bonds, which ensures chemical validity and facilitates the incorporation of structural constraints. We present GraphXForm, a decoder-only graph transformer architecture, which is pretrained on existing compounds and then fine-tuned using a new training algorithm that combines elements of the deep cross-entropy method and self-improvement learning. We evaluate GraphXForm on various drug design tasks, demonstrating superior objective scores compared to state-of-the-art molecular design approaches. Furthermore, we apply GraphXForm to two solvent design tasks for liquid-liquid extraction, again outperforming alternative methods while flexibly enforcing structural constraints or initiating design from existing molecular structures.

cs.LG↗