Search

Previous work on augmenting large multimodal models (LMMs) for text-to-image (T2I) generation has focused on enriching the input space of in-context learning (ICL). This includes providing a few demonstrations and optimizing image descriptions to be more detailed and logical. However, as demand for more complex and flexible image descriptions grows, enhancing comprehension of input text within the ICL paradigm remains a critical yet underexplored area. In this work, we extend this line of research by constructing parallel multilingual prompts aimed at harnessing the multilingual capabilities of LMMs. More specifically, we translate the input text into several languages and provide the models with both the original text and the translations. Experiments on two LMMs across 3 benchmarks show that our method, PMT2I, achieves superior performance in general, compositional, and fine-grained assessments, especially in human preference alignment. Additionally, with its advantage of generating more diverse images, PMT2I significantly outperforms baseline prompts when incorporated with reranking methods. Our code and parallel multilingual data can be found at https://github.com/takagi97/PMT2I., Comment: Accepted to ICASSP 2025

A long-standing puzzle for methanol electrooxidation is how to achieve a CO-free pathway and accurately understand the origin of electrocatalytic selectivity. Herein, we unequivocally demonstrate that the Bi-modified Pt/C follows a CO-free dominated pathway during alkaline methanol electrooxidation, and unveil the formaldehyde (HCHO) intermediate as a critical factor influencing pathway selectivity. These findings are substantiated by kinetic isotope effects, formate Faradaic efficiency, in situ spectroscopy, ab initio molecular dynamic simulations, and density functional theory calculations. Bi modification significantly increases the HCHO dehydrogenation barrier, which facilitates its desorption and subsequent conversion to the H2COOH- anion at the alkaline interface, intrinsically avoiding CO formation. More specifically, the formation of ensemble sites featuring V-shaped Bi-Pt-Bi configuration inhibits the cleavage of C-H bond, and the weak OH binding energy at Bi adatoms effectively prevents blockage of oxygenated species, allowing such ensemble sites to fulfill their functional role. Our study opens up a novel dimension for designing advanced CO-free catalysts., Comment: 16 pages, 6 figures for main text. 54 pages, 47 figures for supplementary material

As large language models (LLMs) evolve, the increase in model depth and parameter number leads to substantial redundancy. To enhance the efficiency of the attention mechanism, previous works primarily compress the KV cache or group attention heads, while largely overlooking redundancy between layers. Our comprehensive analyses across various LLMs show that highly similar attention patterns persist within most layers. It's intuitive to save the computation by sharing attention weights across layers. However, further analysis reveals two challenges: (1) Directly sharing the weight matrix without carefully rearranging the attention heads proves to be ineffective; (2) Shallow layers are vulnerable to small deviations in attention weights. Driven by these insights, we introduce LiSA, a lightweight substitute for self-attention in well-trained LLMs. LiSA employs tiny feed-forward networks to align attention heads between adjacent layers and low-rank matrices to approximate differences in layer-wise attention weights. Evaluations encompassing 13 typical benchmarks demonstrate that LiSA maintains high response quality in terms of accuracy and perplexity while reducing redundant attention calculations within 53-84% of the total layers. Our implementations of LiSA achieve a 6X compression of Q and K, with maximum throughput improvements of 19.5% for LLaMA3-8B and 32.3% for LLaMA2-7B., Comment: Working in process

It was shown in literature that speech representations extracted by self-supervised pre-trained models exhibit similarities with brain activations of human for speech perception and fine-tuning speech representation models on downstream tasks can further improve the similarity. However, it still remains unclear if this similarity can be used to optimize the pre-trained speech models. In this work, we therefore propose to use the brain activations recorded by fMRI to refine the often-used wav2vec2.0 model by aligning model representations toward human neural responses. Experimental results on SUPERB reveal that this operation is beneficial for several downstream tasks, e.g., speaker verification, automatic speech recognition, intent classification.One can then consider the proposed method as a new alternative to improve self-supervised speech models., Comment: accpeted by Interspeech2024