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 1,441 records · Page 80Linked to original sources

arXiv2023

Sumformer: Universal Approximation for Efficient Transformers

Natural language processing (NLP) made an impressive jump with the introduction of Transformers. ChatGPT is one of the most famous examples, changing the perception of the possibilities of AI even outside the research community. However, besides the impressive performance, the quadratic time and space complexity of Transformers with respect to sequence length pose significant limitations for handling long sequences. While efficient Transformer architectures like Linformer and Performer with linear complexity have emerged as promising solutions, their theoretical understanding remains limited. In this paper, we introduce Sumformer, a novel and simple architecture capable of universally approximating equivariant sequence-to-sequence functions. We use Sumformer to give the first universal approximation results for Linformer and Performer. Moreover, we derive a new proof for Transformers, showing that just one attention layer is sufficient for universal approximation.

cs.LG↗

arXiv2024

AttentionLego: An Open-Source Building Block For Spatially-Scalable Large Language Model Accelerator With Processing-In-Memory Technology

Large language models (LLMs) with Transformer architectures have become phenomenal in natural language processing, multimodal generative artificial intelligence, and agent-oriented artificial intelligence. The self-attention module is the most dominating sub-structure inside Transformer-based LLMs. Computation using general-purpose graphics processing units (GPUs) inflicts reckless demand for I/O bandwidth for transferring intermediate calculation results between memories and processing units. To tackle this challenge, this work develops a fully customized vanilla self-attention accelerator, AttentionLego, as the basic building block for constructing spatially expandable LLM processors. AttentionLego provides basic implementation with fully-customized digital logic incorporating Processing-In-Memory (PIM) technology. It is based on PIM-based matrix-vector multiplication and look-up table-based Softmax design. The open-source code is available online: https://bonany.cc/attentionleg.

cs.AR↗

arXiv2024

Exploring Extreme Quantization in Spiking Language Models

Despite the growing prevalence of large language model (LLM) architectures, a crucial concern persists regarding their energy and power consumption, which still lags far behind the remarkable energy efficiency of the human brain. Recent strides in spiking language models (LM) and transformer architectures aim to address this concern by harnessing the spiking activity of biological neurons to enhance energy/power efficiency. Doubling down on the principles of model quantization and energy efficiency, this paper proposes the development of a novel binary/ternary (1/1.58-bit) spiking LM architecture. Achieving scalability comparable to a deep spiking LM architecture is facilitated by an efficient knowledge distillation technique, wherein knowledge from a non-spiking full-precision "teacher" model is transferred to an extremely weight quantized spiking "student" LM. Our proposed model represents a significant advancement as the first-of-its-kind 1/1.58-bit spiking LM, and its performance is rigorously evaluated on multiple text classification tasks of the GLUE benchmark.

cs.NE↗

arXiv2024

Towards a Deeper Understanding of Transformer for Residential Non-intrusive Load Monitoring

Transformer models have demonstrated impressive performance in Non-Intrusive Load Monitoring (NILM) applications in recent years. Despite their success, existing studies have not thoroughly examined the impact of various hyper-parameters on model performance, which is crucial for advancing high-performing transformer models. In this work, a comprehensive series of experiments have been conducted to analyze the influence of these hyper-parameters in the context of residential NILM. This study delves into the effects of the number of hidden dimensions in the attention layer, the number of attention layers, the number of attention heads, and the dropout ratio on transformer performance. Furthermore, the role of the masking ratio has explored in BERT-style transformer training, providing a detailed investigation into its impact on NILM tasks. Based on these experiments, the optimal hyper-parameters have been selected and used them to train a transformer model, which surpasses the performance of existing models. The experimental findings offer valuable insights and guidelines for optimizing transformer architectures, aiming to enhance their effectiveness and efficiency in NILM applications. It is expected that this work will serve as a foundation for future research and development of more robust and capable transformer models for NILM.

eess.SY↗

arXiv2025

Emergent Stack Representations in Modeling Counter Languages Using Transformers

Transformer architectures are the backbone of most modern language models, but understanding the inner workings of these models still largely remains an open problem. One way that research in the past has tackled this problem is by isolating the learning capabilities of these architectures by training them over well-understood classes of formal languages. We extend this literature by analyzing models trained over counter languages, which can be modeled using counter variables. We train transformer models on 4 counter languages, and equivalently formulate these languages using stacks, whose depths can be understood as the counter values. We then probe their internal representations for stack depths at each input token to show that these models when trained as next token predictors learn stack-like representations. This brings us closer to understanding the algorithmic details of how transformers learn languages and helps in circuit discovery.

cs.CL↗

arXiv2025

S2AFormer: Strip Self-Attention for Efficient Vision Transformer

Vision Transformer (ViT) has made significant advancements in computer vision, thanks to its token mixer's sophisticated ability to capture global dependencies between all tokens. However, the quadratic growth in computational demands as the number of tokens increases limits its practical efficiency. Although recent methods have combined the strengths of convolutions and self-attention to achieve better trade-offs, the expensive pairwise token affinity and complex matrix operations inherent in self-attention remain a bottleneck. To address this challenge, we propose S2AFormer, an efficient Vision Transformer architecture featuring novel Strip Self-Attention (SSA). We design simple yet effective Hybrid Perception Blocks (HPBs) to effectively integrate the local perception capabilities of CNNs with the global context modeling of Transformer's attention mechanisms. A key innovation of SSA lies in its reduction of the spatial dimensions of $K$ and $V$, while compressing the channel dimensions of $Q$ and $K$. This design significantly reduces computational overhead while preserving accuracy, striking an optimal balance between efficiency and effectiveness. We evaluate the robustness and efficiency of S2AFormer through extensive experiments on multiple vision benchmarks, including ImageNet-1k for image classification, ADE20k for semantic segmentation, and COCO for object detection and instance segmentation. Results demonstrate that S2AFormer achieves significant accuracy gains with superior efficiency in both GPU and non-GPU environments, making it a strong candidate for efficient vision Transformers.

cs.CV↗

arXiv2025

FTCFormer: Fuzzy Token Clustering Transformer for Image Classification

Transformer-based deep neural networks have achieved remarkable success across various computer vision tasks, largely attributed to their long-range self-attention mechanism and scalability. However, most transformer architectures embed images into uniform, grid-based vision tokens, neglecting the underlying semantic meanings of image regions, resulting in suboptimal feature representations. To address this issue, we propose Fuzzy Token Clustering Transformer (FTCFormer), which incorporates a novel clustering-based downsampling module to dynamically generate vision tokens based on the semantic meanings instead of spatial positions. It allocates fewer tokens to less informative regions and more to represent semantically important regions, regardless of their spatial adjacency or shape irregularity. To further enhance feature extraction and representation, we propose a Density Peak Clustering-Fuzzy K-Nearest Neighbor (DPC-FKNN) mechanism for clustering center determination, a Spatial Connectivity Score (SCS) for token assignment, and a channel-wise merging (Cmerge) strategy for token merging. Extensive experiments on 32 datasets across diverse domains validate the effectiveness of FTCFormer on image classification, showing consistent improvements over the TCFormer baseline, achieving gains of improving 1.43% on five fine-grained datasets, 1.09% on six natural image datasets, 0.97% on three medical datasets and 0.55% on four remote sensing datasets. The code is available at: https://github.com/BaoBao0926/FTCFormer/tree/main.

cs.CV↗

arXiv2025

Foundation Models for Clean Energy Forecasting: A Comprehensive Review

As global energy systems transit to clean energy, accurate renewable generation and renewable demand forecasting is imperative for effective grid management. Foundation Models (FMs) can help improve forecasting of renewable generation and demand because FMs can rapidly process complex, high-dimensional time-series data. This review paper focuses on FMs in the realm of renewable energy forecasting, primarily focusing on wind and solar. We present an overview of the architectures, pretraining strategies, finetuning methods, and types of data used in the context of renewable energy forecasting. We emphasize the role of models that are trained at a large scale, domain specific Transformer architectures, where attention is paid to spatial temporal correlations, the embedding of domain knowledge, and also the brief and intermittent nature of renewable generation. We assess recent FM based advancements in forecast accuracy such as reconciling predictions over multiple time scales and quantifying uncertainty in renewable energy forecasting. We also review existing challenges and areas of improvement in long-term and multivariate time series forecasting. In this survey, a distinction between theory and practice is established regarding the use of FMs in the clean energy forecasting domain. Additionally, it critically assesses the strengths and weaknesses of FMs while advancing future research direction in this new and exciting area of forecasting.

eess.SY↗

arXiv2025

Learning Optimal Crew Dispatch for Grid Restoration Following an Earthquake

Post-disaster crew dispatch is a critical but computationally intensive task. Traditional mixed-integer linear programming methods often require minutes to several hours to compute solutions, leading to delays that hinder timely decision-making in highly dynamic restoration environments. To address this challenge, we propose a novel learning-based framework that integrates transformer architectures with deep reinforcement learning (DRL) to deliver near real-time decision support without compromising solution quality. Crew dispatch is formulated as a sequential decision-making problem under uncertainty, where transformers capture high-dimensional system states and temporal dependencies, while DRL enables adaptive and scalable decision-making. Earthquake-induced distribution network damage is first characterized using established seismic standards, followed by a scenario generation and reduction pipeline that aggregates probable outcomes into a single geospatial impact map. Conditioned on this map, the proposed framework generates second-level dispatch strategies, trained offline on simulated and historical events and deployed online for rapid response. In addition to substantial runtime improvements, the proposed method enhances system resilience by enabling faster and more effective recovery and restoration. Case studies, particularly on the 2869-bus European gas and power network, demonstrate that the method substantially accelerates restoration while maintaining high-quality solutions, underscoring its potential for practical deployment in large-scale disaster response.

eess.SY↗

arXiv2025

Too Helpful, Too Harmless, Too Honest or Just Right?

Large Language Models (LLMs) exhibit strong performance across a wide range of NLP tasks, yet aligning their outputs with the principles of Helpfulness, Harmlessness, and Honesty (HHH) remains a persistent challenge. Existing methods often optimize for individual alignment dimensions in isolation, leading to trade-offs and inconsistent behavior. While Mixture-of-Experts (MoE) architectures offer modularity, they suffer from poorly calibrated routing, limiting their effectiveness in alignment tasks. We propose TrinityX, a modular alignment framework that incorporates a Mixture of Calibrated Experts (MoCaE) within the Transformer architecture. TrinityX leverages separately trained experts for each HHH dimension, integrating their outputs through a calibrated, task-adaptive routing mechanism that combines expert signals into a unified, alignment-aware representation. Extensive experiments on three standard alignment benchmarks-Alpaca (Helpfulness), BeaverTails (Harmlessness), and TruthfulQA (Honesty)-demonstrate that TrinityX outperforms strong baselines, achieving relative improvements of 32.5% in win rate, 33.9% in safety score, and 28.4% in truthfulness. In addition, TrinityX reduces memory usage and inference latency by over 40% compared to prior MoE-based approaches. Ablation studies highlight the importance of calibrated routing, and cross-model evaluations confirm TrinityX's generalization across diverse LLM backbones.

cs.CL↗

arXiv2025

Meta-Learning for Cross-Task Generalization in Protein Mutation Property Prediction

Protein mutations can have profound effects on biological function, making accurate prediction of property changes critical for drug discovery, protein engineering, and precision medicine. Current approaches rely on fine-tuning protein-specific transformers for individual datasets, but struggle with cross-dataset generalization due to heterogeneous experimental conditions and limited target domain data. We introduce two key innovations: (1) the first application of Model-Agnostic Meta-Learning (MAML) to protein mutation property prediction, and (2) a novel mutation encoding strategy using separator tokens to directly incorporate mutations into sequence context. We build upon transformer architectures integrating them with MAML to enable rapid adaptation to new tasks through minimal gradient steps rather than learning dataset-specific patterns. Our mutation encoding addresses the critical limitation where standard transformers treat mutation positions as unknown tokens, significantly degrading performance. Evaluation across three diverse protein mutation datasets (functional fitness, thermal stability, and solubility) demonstrates significant advantages over traditional fine-tuning. In cross-task evaluation, our meta-learning approach achieves 29% better accuracy for functional fitness with 65% less training time, and 94% better accuracy for solubility with 55% faster training. The framework maintains consistent training efficiency regardless of dataset size, making it particularly valuable for industrial applications and early-stage protein design where experimental data is limited. This work establishes a systematic application of meta-learning to protein mutation analysis and introduces an effective mutation encoding strategy, offering transformative methodology for cross-domain generalization in protein engineering.

cs.LG↗

arXiv2026

SeamGen: Artist-Aligned UV Seam Generation via Graph Flow Matching

UV seam placement is a critical yet labor-intensive step in 3D content creation, requiring artists to balance chart shape, seam concealment, and alignment with semantic and geometric features. Existing automatic methods are primarily based on per-object optimization, relying on handcrafted objectives to avoid distortion or on proxies from pretrained models to inject semantic information. However, these strategies are not always well aligned with seams used in industrial production pipelines, often resulting in layouts that deviate from artist-preferred seam patterns and practical production requirements. To address these limitations, we propose SeamGen, a generative model for UV seam generation that aligns with artist preferences and production requirements. Instead of depending on manually designed objectives and constraints, SeamGen learns the distribution of per-edge seam labels from a large corpus of existing seam layouts using a flow-matching generative model. A key challenge is that typical Transformer architectures used in flow matching models are designed for sequential representations, such as point clouds, and cannot naturally account for mesh topology. To enable mesh-native learning, we design a Mesh Transformer backbone that interleaves local graph attention over mesh edges with global self-attention across vertices, capturing both fine-grained geometric cues and long-range topological coherence. To further improve inference-time controllability and quality, we exploit the training-free inpainting capability of flow models for both localized seam refinement and constraint-guided seam generation. Extensive experiments show that by learning priors from professional seam layout data, SeamGen produces UV layouts that better align with artist-authored preferences and achieve superior perceptual quality compared with distortion-based and semantic-proxy baselines.

cs.CV↗

arXiv2025

Robustness Tokens: Towards Adversarial Robustness of Transformers

Recently, large pre-trained foundation models have become widely adopted by machine learning practitioners for a multitude of tasks. Given that such models are publicly available, relying on their use as backbone models for downstream tasks might result in high vulnerability to adversarial attacks crafted with the same public model. In this work, we propose Robustness Tokens, a novel approach specific to the transformer architecture that fine-tunes a few additional private tokens with low computational requirements instead of tuning model parameters as done in traditional adversarial training. We show that Robustness Tokens make Vision Transformer models significantly more robust to white-box adversarial attacks while also retaining the original downstream performances.

cs.LG↗

arXiv2026

UtVAA: Ultra-tiny Vision Transformer with Affix Attention for Mobile Image Classification

Vision Transformers (ViTs) have demonstrated strong representation capability in image classification. However, their quadratic self-attention complexity and large parameter counts limit deployment on resource-constrained mobile and edge devices. This paper introduces UtVAA, an ultra-tiny Vision Transformer architecture designed for efficient visual recognition under strict computational budgets. It incorporates a novel Affix Attention block that combines depthwise-pointwise local feature extraction, linear self-attention, coordinate attention for spatial dependency modelling, and a lightweight ternary fusion strategy to integrate local and global representations. In addition, Dilated Bottleneck blocks expand the receptive field using dilated depthwise separable convolutions while maintaining low FLOPs and stable optimisation through residual connections. UtVAA is implemented in scalable Tiny, Medium, and Large variants, with the smallest model containing 204.67K parameters and 53.95M FLOPs. Experimental results on CIFAR-10, CIFAR-100, PlantVillage-Tomato and SLIF-Tomato datasets show that UtVAA achieves competitive accuracy within a sub-million-parameter regime. Overall, the results demonstrate that transformer-based vision models can be redesigned into ultra-tiny architectures without significant loss in discriminative performance, making UtVAA suitable for mobile and edge deployment. Code is available at https://github.com/romiyal/UtVAA

cs.CV↗

arXiv2020

Regularized Context Gates on Transformer for Machine Translation

Context gates are effective to control the contributions from the source and target contexts in the recurrent neural network (RNN) based neural machine translation (NMT). However, it is challenging to extend them into the advanced Transformer architecture, which is more complicated than RNN. This paper first provides a method to identify source and target contexts and then introduce a gate mechanism to control the source and target contributions in Transformer. In addition, to further reduce the bias problem in the gate mechanism, this paper proposes a regularization method to guide the learning of the gates with supervision automatically generated using pointwise mutual information. Extensive experiments on 4 translation datasets demonstrate that the proposed model obtains an averaged gain of 1.0 BLEU score over a strong Transformer baseline.

cs.CL↗

arXiv2020

Multi-modal Transformer for Video Retrieval

The task of retrieving video content relevant to natural language queries plays a critical role in effectively handling internet-scale datasets. Most of the existing methods for this caption-to-video retrieval problem do not fully exploit cross-modal cues present in video. Furthermore, they aggregate per-frame visual features with limited or no temporal information. In this paper, we present a multi-modal transformer to jointly encode the different modalities in video, which allows each of them to attend to the others. The transformer architecture is also leveraged to encode and model the temporal information. On the natural language side, we investigate the best practices to jointly optimize the language embedding together with the multi-modal transformer. This novel framework allows us to establish state-of-the-art results for video retrieval on three datasets. More details are available at http://thoth.inrialpes.fr/research/MMT.

cs.CV↗

arXiv2022

Learning Versatile Neural Architectures by Propagating Network Codes

This work explores how to design a single neural network capable of adapting to multiple heterogeneous vision tasks, such as image segmentation, 3D detection, and video recognition. This goal is challenging because both network architecture search (NAS) spaces and methods in different tasks are inconsistent. We solve this challenge from both sides. We first introduce a unified design space for multiple tasks and build a multitask NAS benchmark (NAS-Bench-MR) on many widely used datasets, including ImageNet, Cityscapes, KITTI, and HMDB51. We further propose Network Coding Propagation (NCP), which back-propagates gradients of neural predictors to directly update architecture codes along the desired gradient directions to solve various tasks. In this way, optimal architecture configurations can be found by NCP in our large search space in seconds. Unlike prior arts of NAS that typically focus on a single task, NCP has several unique benefits. (1) NCP transforms architecture optimization from data-driven to architecture-driven, enabling joint search an architecture among multitasks with different data distributions. (2) NCP learns from network codes but not original data, enabling it to update the architecture efficiently across datasets. (3) In addition to our NAS-Bench-MR, NCP performs well on other NAS benchmarks, such as NAS-Bench-201. (4) Thorough studies of NCP on inter-, cross-, and intra-tasks highlight the importance of cross-task neural architecture design, i.e., multitask neural architectures and architecture transferring between different tasks. Code is available at https://github.com/dingmyu/NCP.

cs.CV↗

arXiv2021

Beyond Isolated Utterances: Conversational Emotion Recognition

Speech emotion recognition is the task of recognizing the speaker's emotional state given a recording of their utterance. While most of the current approaches focus on inferring emotion from isolated utterances, we argue that this is not sufficient to achieve conversational emotion recognition (CER) which deals with recognizing emotions in conversations. In this work, we propose several approaches for CER by treating it as a sequence labeling task. We investigated transformer architecture for CER and, compared it with ResNet-34 and BiLSTM architectures in both contextual and context-less scenarios using IEMOCAP corpus. Based on the inner workings of the self-attention mechanism, we proposed DiverseCatAugment (DCA), an augmentation scheme, which improved the transformer model performance by an absolute 3.3% micro-f1 on conversations and 3.6% on isolated utterances. We further enhanced the performance by introducing an interlocutor-aware transformer model where we learn a dictionary of interlocutor index embeddings to exploit diarized conversations.

cs.CL↗