Papers with computation overhead
Mask What Matters: Mitigating Object Hallucinations in Multimodal Large Language Models with Object-Aligned Visual Contrastive Decoding (2026.eacl-srw)
Copied to clipboard
| Challenge: | Recent studies improve visual contrastive decoding (VCD) by constructing more informative auxiliary views. |
| Approach: | They propose to construct an object-aligned auxiliary view that disrupts unsupported tokens and produces a stronger contrast signal. |
| Outcome: | Empirically, the proposed method shows consistent gains on two popular object hallucination benchmarks across two MLLMs. |
Routing to the Expert: Efficient Reward-guided Ensemble of Large Language Models (2024.naacl-long)
Copied to clipboard
| Challenge: | Existing ensemble methods for Large Language Models focus on reward model ranking of outputs, leading to significant computation overhead. |
| Approach: | They propose a reward-guided routing method distilling rewards on training queries to train a routing function. |
| Outcome: | The proposed method outperforms the best single model and ranks first on 44% of tasks. |
LEALLA: Learning Lightweight Language-agnostic Sentence Embeddings with Knowledge Distillation (2023.eacl-main)
Copied to clipboard
| Challenge: | Large-scale language-agnostic sentence embedding models suffer from inference speed and computation overhead. |
| Approach: | They propose to train a lightweight sentence embedding model to achieve this by incorporating knowledge from a teacher model. |
| Outcome: | The proposed model can build low-dimensional sentences for 109 languages with a thin-deep encoder. |
When in Doubt: Improving Classification Performance with Alternating Normalization (2021.findings-emnlp)
Copied to clipboard
| Challenge: | a classifier that uses a nonparametric post-processing step for classification suffers when given examples that are close to its decision boundary. |
| Approach: | They propose a nonparametric post-processing step that re-adjusts predicted class probability distributions using high-confidence validation examples. |
| Outcome: | The proposed method improves classifier accuracy on difficult examples. |
InfiniSST: Simultaneous Translation of Unbounded Speech with Large Language Model (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing models for simultaneous speech translation assume pre-segmented speech, limiting their real-world applicability. |
| Approach: | They propose a multi-turn dialogue task that can translate unbounded streaming speech . they construct translation trajectories and robust segments from MuST-C with multi-latency augmentation during training and develop a cache management strategy to facilitate efficient inference. |
| Outcome: | The proposed approach reduces computation-aware latency by 0.5 to 1 second while maintaining the same translation quality compared to baselines. |
Active2 Learning: Actively reducing redundancies in Active Learning methods for Sequence Tagging and Machine Translation (2021.naacl-main)
Copied to clipboard
| Challenge: | Existing approaches to deep learning for NLP require large amounts of labeled data. |
| Approach: | They propose an approach that iteratively selects a small number of examples for expert annotation based on their estimated utility in training the model. |
| Outcome: | The proposed approach reduces the data requirements of state-of-the-art AL strategies by 3-25% on multiple NLP tasks while achieving the same performance with virtually no additional computation overhead. |
ProFormer: Towards On-Device LSH Projection Based Transformers (2021.eacl-main)
Copied to clipboard
| Challenge: | ProFormer is a projection based transformer architecture that is faster and lighter making it suitable to deploy to memory constraint devices such as mobile phones, watches and IoT. |
| Approach: | They propose a projection based transformer architecture that generates word representations on-the-fly without embedding lookup tables and a local projection attention layer that transforms the input sequence of N LSH word projections into a sequence of K representations. |
| Outcome: | The proposed architecture reduces memory footprint from 92.16 MB to 1.7 KB and requires 16x less computation overhead making it suitable to deploy to memory constraint devices and preserve user privacy. |
Length-Adaptive Distillation: Customizing Small Language Model for Dynamic Token Pruning (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods to accelerate inference speed are model compression and dynamic computation (e.g., dynamic token pruning). |
| Approach: | They propose a two-stage knowledge distillation framework that produces a customized small language model for dynamic token pruning. |
| Outcome: | The proposed framework can make the small language model more customized for dynamic token pruning and achieve better speed-performance trade-off. |
TurboRAG: Accelerating Retrieval-Augmented Generation with Precomputed KV Caches for Chunked Text (2025.emnlp-main)
Copied to clipboard
| Challenge: | Current RAG systems concatenate and process numerous retrieved document chunks for prefill . this leads to significant latency in time-to-first-token (TTFT) Experimental results demonstrate that TurboRAG reduces TTFT by up to 9.4x compared to the conventional RAG system. |
| Approach: | They propose a hybrid offline-online paradigm that precomputes chunk-level key-value caches and stitches them together at inference time using independent–attention and reorderedRoPE techniques. |
| Outcome: | Experimental results show that TurboRAG reduces TTFT by 9.4x compared to the conventional RAG systems . long concatenated contexts consume disproportionate GPU memory, limiting throughput . |
CoMoE: Contrastive Representation for Mixture-of-Experts in Parameter-Efficient Fine-tuning (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Currently, mixture-of-experts (MoE) is underutilized on heterogeneous datasets, ignoring the fact that experts may learn similar knowledge. |
| Approach: | They propose a method to promote modularization and specialization in MoE by specializing functionalities into different experts and sparsely activating them appropriately. |
| Outcome: | The proposed method improves the capacity and specialization of mixture-of-experts (MoE) by sampling from activated and inactivated experts in top-k routing. |
LaMemo: Language Modeling with Look-Ahead Memory (2022.naacl-main)
Copied to clipboard
| Challenge: | Existing approaches to model long-term dependencies are limited to long texts with thousands of words. |
| Approach: | They propose a look-ahead memory that augments the recurrence memory by attending to the right-side tokens and interpolating with the old memory states to maintain long-term information in the history. |
| Outcome: | Experiments on widely used language modeling benchmarks show that LaMemo outperforms baseline models with recurrence memory. |
Powerformer: Efficient and High-Accuracy Privacy-Preserving Language Model with Homomorphic Encryption (2025.acl-long)
Copied to clipboard
| Challenge: | a new privacy-preserving language model, Powerformer, is designed to reduce computation overhead while maintaining model performance. |
| Approach: | They propose an efficient homomorphic encryption-based privacy-preserving language model . it incorporates three key techniques to optimize encrypted computations . |
| Outcome: | The proposed model achieves 45% reduction in computation time compared to state-of-the-art models . authors say the model preserves data privacy and AI capabilities in MLaaS environments . |
Language Detoxification with Attribute-Discriminative Latent Space (2023.acl-long)
Copied to clipboard
| Challenge: | Existing methods to detoxify toxic text require excessive memory, computations and time. |
| Approach: | They propose a method to generate toxic text using an attribute-discriminative latent space. |
| Outcome: | The proposed method outperforms baselines on detoxified language and dialogue generation tasks while being time- and memory-efficient. |
Prompting ELECTRA: Few-Shot Learning with Discriminative Pre-Trained Models (2022.emnlp-main)
Copied to clipboard
| Challenge: | Pre-trained masked language models perform few-shot learning, but discriminative models like ELECTRA do not fit into the paradigm. |
| Approach: | They propose to use ELECTRA to train pre-trained models to score originality of target options without introducing new parameters. |
| Outcome: | The proposed model outperforms masked language models in a wide range of tasks without adding new parameters. |
Extractive Medical Entity Disambiguation with Memory Mechanism and Memorized Entity Information (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods focus on local optimal while ignoring sole-mention disambiguation boosted by richer context from other mentions’ disambiguating processes. |
| Approach: | They propose an approach to extracting medical entity disambiguation using memory mechanism and memorized entity information (M3E) they use a memory mechanism module that performs memory caching, retrieval, fusion and cross-network residual to aid the disambiguations of remaining mentions. |
| Outcome: | The proposed method outperforms state-of-the-art methods on two benchmark datasets. |