Automated Audio Captioning (AAC) is a multimodal task aimed at generating natural language descriptions of audio content. Previous studies have shown that LLMs can improve AAC performance by summarizing audio events based on a list of candidate captions, which are selected by an external reranker from those generated using Nucleus Sampling. However, the reranking process often selects overly similar captions, disregarding the original diversity of the sampled captions. In this work, we show that this diversity reflects the AAC model’s level of certainty and propose a lightweight candidate selection approach that preserves the initial diversity of the generated captions. This, in turn, enables an LLM to summarize the captions while considering the AAC model’s certainty in a few-shot setting. Experimental results demonstrate that our method outperforms previous post-processing techniques while being significantly faster.

This paper presents our submission to the Iranian division of the Text-Dependent Speaker Verification Challenge (TdSV) 2024. Conventional TdSV approaches typically jointly model speaker and linguistic features, requiring unsegmented inputs during training and incurring high computational costs. Additionally, these methods often fine-tune large-scale pre-trained speaker embedding models on the target domain dataset, which may compromise the pre-trained models’ original ability to capture speaker-specific characteristics. To overcome these limitations, we employ a TdSV system that utilizes two pre-trained models independently and demonstrate that, by leveraging pre-trained models with targeted domain adaptation, competitive results can be achieved while avoiding the substantial computational costs associated with joint fine-tuning on unsegmented inputs in conventional approaches. Our best system reached a MinDCF of 0.0358 on the evaluation subset and secured first place in the challenge.

Recent advancements in natural language processing (NLP) have prompted the development of sophisticated reasoning benchmarks. This paper presents our system for the SemEval 2024 Task 9 competition and also investigates the efficacy of fine-tuning language models (LMs) on BrainTeaser—a benchmark designed to evaluate NLP models’ lateral thinking and creative reasoning abilities. Our experiments focus on two prominent families of pre-trained models, BERT and T5. Additionally, we explore the potential benefits of multi-task fine-tuning on commonsense reasoning datasets to enhance performance. Our top-performing model, DeBERTa-v3-large, achieves an impressive overall accuracy of 93.33%, surpassing human performance.