Stickers, while widely recognized for enhancing empathetic communication in online interactions, remain underexplored in current empathetic dialogue research. In this paper, we introduce the Agent for StickerConv (Agent4SC), which uses collaborative agent interactions to realistically simulate human behavior with sticker usage, thereby enhancing multimodal empathetic communication. Building on this foundation, we develop a multimodal empathetic dialogue dataset, StickerConv, which includes 12.9K dialogue sessions, 5.8K unique stickers, and 2K diverse conversational scenarios, specifically designs to augment the generation of empathetic responses in a multimodal context. To leverage the richness of this dataset, we propose PErceive and Generate Stickers (PEGS), a multimodal empathetic response generation model, complemented by a comprehensive set of empathy evaluation metrics based on LLM. Our experiments demonstrate PEGS's effectiveness in generating contextually relevant and emotionally resonant multimodal empathetic responses, contributing to the advancement of more nuanced and engaging empathetic dialogue systems. Our project page is available at https://neu-datamining.github.io/StickerConv .

In this study, we introduce Orion-14B, a collection of multilingual large language models with 14 billion parameters. We utilize a data scheduling approach to train a foundational model on a diverse corpus of 2.5 trillion tokens, sourced from texts in English, Chinese, Japanese, Korean, and other languages. Additionally, we fine-tuned a series of models tailored for conversational applications and other specific use cases. Our evaluation results demonstrate that Orion-14B achieves state-of-the-art performance across a broad spectrum of tasks.

In the face of rapidly expanding online medical literature, automated systems for aggregating and summarizing information are becoming increasingly crucial for healthcare professionals and patients. Large Language Models (LLMs), with their advanced generative capabilities, have shown promise in various NLP tasks, and their potential in the healthcare domain, particularly for Closed-Book Generative QnA, is significant. However, the performance of these models in domain-specific tasks such as medical Q&A remains largely unexplored. This study aims to fill this gap by comparing the performance of general and medical-specific distilled LMs for medical Q&A. We aim to evaluate the effectiveness of fine-tuning domain-specific LMs and compare the performance of different families of Language Models. The study will address critical questions about these models' reliability, comparative performance, and effectiveness in the context of medical Q&A. The findings will provide valuable insights into the suitability of different LMs for specific applications in the medical domain.

Mapping speech tokens to the same feature space as text tokens has become the paradigm for the integration of speech modality into decoder-only large language models (LLMs). An alternative approach is to use an encoder-decoder architecture that incorporates speech features through cross-attention. This approach, however, has received less attention in the literature. In this work, we connect the Whisper encoder with ChatGLM3 and provide in-depth comparisons of these two approaches using Chinese automatic speech recognition (ASR) and name entity recognition (NER) tasks. We evaluate them not only by conventional metrics like the F1 score but also by a novel fine-grained taxonomy of ASR-NER errors. Our experiments reveal that encoder-decoder architecture outperforms decoder-only architecture with a short context, while decoder-only architecture benefits from a long context as it fully exploits all layers of the LLM. By using LLM, we significantly reduced the entity omission errors and improved the entity ASR accuracy compared to the Conformer baseline. Additionally, we obtained a state-of-the-art (SOTA) F1 score of 0.805 on the AISHELL-NER test set by using chain-of-thought (CoT) NER which first infers long-form ASR transcriptions and then predicts NER labels.

Interpreting hierarchical structures latent in language is a key limitation of current language models (LMs). While previous research has implicitly leveraged these hierarchies to enhance LMs, approaches for their explicit encoding are yet to be explored. To address this, we introduce a novel approach to re-train transformer encoder-based LMs as Hierarchy Transformer encoders (HiTs), harnessing the expansive nature of hyperbolic space. Our method situates the output embedding space of pre-trained LMs within a Poincar\'e ball with a curvature that adapts to the embedding dimension, followed by re-training on hyperbolic cluster and centripetal losses. These losses are designed to effectively cluster related entities (input as texts) and organise them hierarchically. We evaluate HiTs against pre-trained and fine-tuned LMs, focusing on their capabilities in simulating transitive inference, predicting subsumptions, and transferring knowledge across hierarchies. The results demonstrate that HiTs consistently outperform both pre-trained and fine-tuned LMs in these tasks, underscoring the effectiveness and transferability of our re-trained hierarchy encoders.

Vision-language models such as CLIP have shown impressive capabilities in encoding texts and images into aligned embeddings, enabling the retrieval of multimodal data in a shared embedding space. However, these embedding-based models still face challenges in effectively matching images and texts with similar visio-linguistic compositionality, as evidenced by their performance on the recent Winoground dataset. In this paper, we argue that this limitation stems from two factors: the use of single vector representations for complex multimodal data, and the absence of step-by-step reasoning in these embedding-based methods. To address this issue, we make an exploratory step using a novel generative method that prompts large vision-language models (e.g., GPT-4) to depict images and perform compositional reasoning. Our method outperforms other embedding-based methods on the Winoground dataset, and obtains further improvement of up to 10% accuracy when enhanced with the optimal description.

In-context learning (ICL) has become an effective solution for few-shot learning in natural language processing. Past work has found that, during this process, representations of the last prompt token are utilized to store task reasoning procedures, thereby explaining the working mechanism of in-context learning. In this paper, we seek to locate and analyze other task-encoding tokens whose representations store task reasoning procedures. Supported by experiments that ablate the representations of different token types, we find that template and stopword tokens are the most prone to be task-encoding tokens. In addition, we demonstrate experimentally that lexical cues, repetition, and text formats are the main distinguishing characteristics of these tokens. Our work provides additional insights into how large language models (LLMs) leverage task reasoning procedures in ICL and suggests that future work may involve using task-encoding tokens to improve the computational efficiency of LLMs at inference time and their ability to handle long sequences.

With the proliferation of large pre-trained language models (PLMs), fine-tuning all model parameters becomes increasingly inefficient, particularly when dealing with numerous downstream tasks that entail substantial training and storage costs. Several approaches aimed at achieving parameter-efficient fine-tuning (PEFT) have been proposed. Among them, Low-Rank Adaptation (LoRA) stands out as an archetypal method, incorporating trainable rank decomposition matrices into each target module. Nevertheless, LoRA does not consider the varying importance of each layer. To address these challenges, we introduce PRILoRA, which linearly allocates a different rank for each layer, in an increasing manner, and performs pruning throughout the training process, considering both the temporary magnitude of weights and the accumulated statistics of the input to any given layer. We validate the effectiveness of PRILoRA through extensive experiments on eight GLUE benchmarks, setting a new state of the art.

Timely, personalized feedback is essential for students learning programming, especially as class sizes expand. LLM-based tools like ChatGPT offer instant support, but reveal direct answers with code, which may hinder deep conceptual engagement. We developed CodeAid, an LLM-based programming assistant delivering helpful, technically correct responses, without revealing code solutions. For example, CodeAid can answer conceptual questions, generate pseudo-code with line-by-line explanations, and annotate student's incorrect code with fix suggestions. We deployed CodeAid in a programming class of 700 students for a 12-week semester. A thematic analysis of 8,000 usages of CodeAid was performed, further enriched by weekly surveys, and 22 student interviews. We then interviewed eight programming educators to gain further insights on CodeAid. Findings revealed students primarily used CodeAid for conceptual understanding and debugging, although a minority tried to obtain direct code. Educators appreciated CodeAid's educational approach, and expressed concerns about occasional incorrect feedback and students defaulting to ChatGPT.

We propose a natural language prompt-based retrieval augmented generation (Prompt-RAG), a novel approach to enhance the performance of generative large language models (LLMs) in niche domains. Conventional RAG methods mostly require vector embeddings, yet the suitability of generic LLM-based embedding representations for specialized domains remains uncertain. To explore and exemplify this point, we compared vector embeddings from Korean Medicine (KM) and Conventional Medicine (CM) documents, finding that KM document embeddings correlated more with token overlaps and less with human-assessed document relatedness, in contrast to CM embeddings. Prompt-RAG, distinct from conventional RAG models, operates without the need for embedding vectors. Its performance was assessed through a Question-Answering (QA) chatbot application, where responses were evaluated for relevance, readability, and informativeness. The results showed that Prompt-RAG outperformed existing models, including ChatGPT and conventional vector embedding-based RAGs, in terms of relevance and informativeness. Despite challenges like content structuring and response latency, the advancements in LLMs are expected to encourage the use of Prompt-RAG, making it a promising tool for other domains in need of RAG methods.