With the rapid advancement of machine translation research, evaluation toolkits have become essential for benchmarking system progress. Tools like COMET and SacreBLEU offer single quality score assessments that are ef...
详细信息
The success of large language models (LLMs), like GPT-4 and ChatGPT, has led to the development of numerous cost-effective and accessible alternatives that are created by finetuning open-access LLMs with task-specific...
详细信息
ISBN:
(纸本)9798891760608
The success of large language models (LLMs), like GPT-4 and ChatGPT, has led to the development of numerous cost-effective and accessible alternatives that are created by finetuning open-access LLMs with task-specific data (e.g., ChatDoctor) or instruction data (e.g., Alpaca). Among the various fine-tuning methods, adapter-based parameter-efficient fine-tuning (PEFT) is undoubtedly one of the most attractive topics, as it only requires fine-tuning a few external parameters instead of the entire LLMs while achieving comparable or even better performance. To enable further research on PEFT methods of LLMs, this paper presents LLM-Adapters, an easy-to-use framework that integrates various adapters into LLMs and can execute these adapter-based PEFT methods of LLMs for different tasks. The framework includes state-of-the-art open-access LLMs such as LLaMA, BLOOM, and GPT-J, as well as widely used adapters such as Series adapters, Parallel adapter, Prompt-based learning and Reparametrization-based methods. Moreover, we conduct extensive empirical studies on the impact of adapter types, placement locations, and hyper-parameters to the best design for each adapter-based methods. We evaluate the effectiveness of the adapters on fourteen datasets from two different reasoning tasks, Arithmetic Reasoning and Commonsense Reasoning. The results demonstrate that using adapter-based PEFT in smaller-scale LLMs (7B) with few extra trainable parameters yields comparable, and in some cases superior, performance to powerful LLMs (175B) in zero-shot inference on both reasoning tasks. The code and datasets can be found in https://***/AGI-Edgerunners/LLM-Adapters.
We present a comprehensive evaluation of large language models for multilingual readability assessment. Existing evaluation resources lack domain and language diversity, limiting the ability for cross-domain and cross...
详细信息
To enhance a question-answering system for automotive drivers, we tackle the problem of automatic generation of icon image descriptions. The descriptions can match the driver’s query about the icon appearing on the d...
详细信息
The common toxicity and societal bias in contents generated by large language models (LLMs) necessitate strategies to reduce harm. Present solutions often demand white-box access to the model or substantial training, ...
详细信息
Large language models (LLMs) have shown significant achievements in solving a wide range of tasks. Recently, LLMs' capability to store, retrieve and infer with symbolic knowledge has drawn a great deal of attentio...
详细信息
Many datasets have been developed to train and evaluate document-level relation extraction (RE) models. Most of these are constructed using real-world data. It has been shown that RE models trained on real-world data ...
详细信息
This systematic review explores the integration and impact of Artificial Intelligence in English as a Foreign language teaching in schools, evaluating the effectiveness, challenges, and pedagogical implications of AI-...
详细信息
This systematic review explores the integration and impact of Artificial Intelligence in English as a Foreign language teaching in schools, evaluating the effectiveness, challenges, and pedagogical implications of AI-driven tools. After screening 189 studies from seven databases, 22 relevant empirical studies focusing on experiential learning outcomes with AI use were selected, following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The findings highlight AI's transformative impact on school-based EFL education, offering tailored, interactive experiences. Students using AI tools showed significant improvements in reading, writing, listening, speaking, vocabulary, and overall language comprehension compared to traditional methods. Improvements in language proficiency align with all three domains of Bloom's Taxonomy. Tools like naturallanguageprocessing and Intelligent Tutoring Systems enhance instruction but struggle with language nuances and cultural contexts. Challenges like the digital divide, literacy gaps, teacher readiness and role confusion, cognitive load, and context-specific adaptation persist. Addressing these requires robust infrastructure, teacher training, and institutional support. The review offers valuable insights for teachers, policymakers, and researchers dedicated to advancing school-based EFL education with innovative AI solutions.
Despite progress in multimodal large language models (MLLMs), the challenge of interpreting long-form videos in response to linguistic queries persists, largely due to the inefficiency in temporal grounding and limite...
详细信息
Autoregressive (AR) encoder-decoder neural networks have proved successful in many NLP problems, including Semantic Parsing - a task that translates naturallanguage to machine-readable parse trees. However, the seque...
详细信息
暂无评论