Mixture-of-Modules: Reinventing Transformers as Dynamic Assemblies of Modules (2024.emnlp-main)
Copied to clipboard
| Challenge: | Empirical results show that MoMs consistently outperform vanilla transformers . |
| Approach: | They propose an architecture that allows for a mixture-of-modules computation that uses a finite set of modules defined by multi-head attention and feed-forward networks. |
| Outcome: | The proposed architecture outperforms vanilla Transformers and their variants in multiple ways. |
Similar Papers
Approximating Two-Layer Feedforward Networks for Efficient Transformers (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Recent work uses sparse Mixtures of Experts (MoEs) to build resource-efficient large language models. |
| Approach: | They propose a general framework that unifies various methods to build two-layer NNs . they propose methods to improve both MoEs and PKMs based on their results . |
| Outcome: | The proposed framework improves both MoEs and product-key memories (PKMs) it shows that MoE's are competitive with dense Transformer-XL on two different scales while being much more resource efficient. |
Token-Wise Kernels (TWiKers) for Vicinity-Aware Attention in Transformers (2026.findings-eacl)
Copied to clipboard
| Challenge: | Token-Wise Kernels (TWiKers) are a novel enhancement to transformers that learn token-specific convolutional kernels applied to the keys or values. |
| Approach: | They propose a transformer enhancement that learns token-specific convolutional kernels applied to the keys or values. |
| Outcome: | The proposed transformers learn token-specific convolutional kernels applied to the keys or values . the results show that content words retain self-focus while function words shift attention toward their neighbors . |
Modular Transformers: Compressing Transformers into Modularized Layers for Flexible Efficient Inference (2023.findings-acl)
Copied to clipboard
| Challenge: | Pre-trained sequence-to-sequence models have advanced the state of the art on text generation tasks. |
| Approach: | They introduce a modular encoder-decoder framework for flexible sequence-to-sequence model compression. |
| Outcome: | The proposed framework can achieve flexible compression ratios from 1.1x to 6x with little to moderate relative performance drop. |
FNet: Mixing Tokens with Fourier Transforms (2022.naacl-main)
Copied to clipboard
| Challenge: | Using simple linear transformations, Transformer encoders can be sped up with limited accuracy costs by replacing the self-attention sublayers with simple linear mixing mechanisms. |
| Approach: | They propose to replace the self-attention sublayer with a linear transformation that "mixes" input tokens. |
| Outcome: | The proposed model outperforms the “efficient Transformers” on the GLUE benchmark at longer input lengths and on smaller models with a light memory footprint. |
MoEfication: Transformer Feed-forward Layers are Mixtures of Experts (2022.findings-acl)
Copied to clipboard
| Challenge: | Recent work has shown that feed-forward networks (FFNs) in pre-trained Transformers are a key component, storing various linguistic and factual knowledge. |
| Approach: | They propose to convert a model into its MoE version with the same parameters and build expert routers to decide which experts will be used for each input. |
| Outcome: | The proposed model can use 10% to 30% of FFN parameters while maintaining over 95% original performance. |
Mixture of Attention Heads: Selecting Attention Heads Per Token (2022.emnlp-main)
Copied to clipboard
| Challenge: | Mixture-of-Experts (MoE) networks have been proposed as an efficient way to scale up model capacity and implement conditional computing. |
| Approach: | They propose a new architecture that combines multi-head attention with the MoE mechanism and a sparsely gated architecture that allows for faster computations. |
| Outcome: | The proposed architecture can scale up the number of attention heads and the number parameters while preserving computational efficiency. |
LAIT: Efficient Multi-Segment Encoding in Transformers with Layer-Adjustable Interaction (2023.acl-long)
Copied to clipboard
Jeremiah Milbauer, Annie Louis, Mohammad Javad Hosseini, Alex Fabrikant, Donald Metzler, Tal Schuster
| Challenge: | In many NLP tasks, the input text can be seen as a sequence of related segments. |
| Approach: | They propose a layer-adjustable interactions framework that contextualizes token representations by attending to all other tokens at each layer, leading to quadratic increase in compute effort with the input length. |
| Outcome: | The proposed model reduces 30-50% of attention FLOPs while maintaining high accuracy. |
Inceptive Transformers: Enhancing Contextual Representations through Multi-Scale Feature Learning Across Domains and Languages (2025.emnlp-main)
Copied to clipboard
| Challenge: | Encoder transformer models compress information from all tokens into a single [CLS] token to represent global context. |
| Approach: | They propose a 1-D convolution module that augments token representations with multi-scale local features to improve performance. |
| Outcome: | Experiments on five diverse tasks show that the proposed framework outperforms baseline models by 1% to 14% while maintaining efficiency. |
Leap-of-Thought: Accelerating Transformers via Dynamic Token Routing (2023.emnlp-main)
Copied to clipboard
| Challenge: | Inefficient transformers have been a challenge for many years, requiring computational costs that scale quadratically with the length of the input sequence. |
| Approach: | They propose a token reduction approach that dynamically routes tokens within layers to ensure that all tokens remain accessible in subsequent layers. |
| Outcome: | The proposed approach achieves up to 25x faster inference time without significant loss in accuracy. |
Transformer Feed-Forward Layers Are Key-Value Memories (2021.emnlp-main)
Copied to clipboard
| Challenge: | Using feed-forward layers, we show that the learned patterns are human-interpretable, and that lower layers tend to capture shallow patterns, while upper layers learn more semantic ones. |
| Approach: | They propose that feed-forward layers in transformer-based language models operate as key-value memories where each key correlates with textual patterns in the training examples and each value induces a distribution over the output vocabulary. |
| Outcome: | The proposed model is based on key-value memories with a key-level correlation with the training examples and a distribution over the output vocabulary. |