Instruction tuning has been attracting much attention to achieve generalization ability across a wide variety of tasks. Although various types of instructions have been manually created for instruction tuning, it is still unclear what kind of instruction is optimal to obtain cross-task generalization ability. This work presents instruction optimization, which optimizes training instructions with respect to generalization ability. Rather than manually tuning instructions, we introduce learnable instructions and optimize them with gradient descent by leveraging bilevel optimization. Experimental results show that the learned instruction enhances the diversity of instructions and improves the generalization ability compared to using only manually created instructions.
Allen Institute for AI（艾伦人工智能研究所）；University of Washington（华盛顿大学）；Arizona State University（亚利桑那州立大学）
Humans (e.g., crowdworkers) have a remarkable ability in solving different tasks, by simply reading textual instructions that define them and looking at a few examples. Despite the success of the conventional supervised learning on individual datasets, such models often struggle with generalization across tasks (e.g., a question-answering system cannot solve classification tasks). A long-standing challenge in AI is to build a model that learns a new task by understanding the human-readable instructions that define it. To study this, we introduce NATURAL INSTRUCTIONS, a dataset of 61 distinct t…
On the Effect of Instruction Tuning Loss on Generalization
Dept. of Electrical Engineering（电气工程系）；Indian Institute of Technology Delhi（印度理工学院德里）；Media and Data Science Research（媒体与数据科学研究）；Adobe Inc., India（Adobe公司，印度）
From Templates to Natural Language: Generalization Challenges in Instruction-Tuned LLMs for Spatial Reasoning
Computational Linguistics, Department of Linguistics University of Potsdam（乌特雷赫特大学语言学系计算语言学部）；German Research Center for Artificial Intelligence (DFKI), Berlin（德国人工智能研究中心（DFKI）柏林）
Tsinghua University（清华大学）；Institute for Artificial Intelligence（人工智能研究院）；Beijing National Research Center for Information Science and Technology（北京信息科学与技术国家研究中心）
Training language models to learn from human instructions for zero-shot cross-task generalization has attracted much attention in NLP communities. Recently, instruction tuning (IT), which fine-tunes a pre-trained language model on a massive collection of tasks described via human-craft instructions, has been shown effective in instruction learning for unseen tasks. However, IT relies on a large amount of human-annotated samples, which restricts its generalization. Unlike labeled data, unlabeled data are often massive and cheap to obtain. In this work, we study how IT can be improved with unlab…
Instruction tuning is an emergent paradigm in NLP wherein natural language instructions are leveraged with language models to induce zero-shot performance on unseen tasks. Instructions have been shown to enable good performance on unseen tasks and datasets in both large and small language models. Dialogue is an especially interesting area to explore instruction tuning because dialogue systems perform multiple kinds of tasks related to language (e.g., natural language understanding and generation, domain-specific interaction), yet instruction tuning has not been systematically explored for dial…
Recent work has shown that fine-tuning large pre-trained language models on a collection of tasks described via instructions, a.k.a. instruction-tuning, improves their zero and few-shot generalization to unseen tasks. However, there is a limited understanding of the performance trade-offs of different decisions made during the instruction-tuning process. These decisions include the scale and diversity of the instruction-tuning benchmark, different task sampling strategies, fine-tuning with and without demonstrations, training using specialized datasets for reasoning and dialogue, and finally, …
From Specialization to Generalization: Instruction-tuned LLMs for Robust Harmful Content Mitigation
School of Computation, Information and Technology, TU Munich（慕尼黑工业大学计算、信息与技术学院）；Munich Center for Machine Learning（慕尼黑机器学习中心）；Munich Data Science Institute（慕尼黑数据科学研究所）
Georgia Institute of Technology（佐治亚理工学院）；Facebook AI Research（脸书人工智能研究院）；Oregon State University（俄勒冈州立大学）
A visually-grounded navigation instruction can be interpreted as a sequence of expected observations and actions an agent following the correct trajectory would encounter and perform. Based on this intuition, we formulate the problem of finding the goal location in Vision-and-Language Navigation (VLN) within the framework of Bayesian state tracking - learning observation and motion models conditioned on these expectable events. Together with a mapper that constructs a semantic spatial map on-the-fly during navigation, we formulate an end-to-end differentiable Bayes filter and train it to ident…
On Conditional and Compositional Language Model Differentiable Prompting
Mila - Québec AI Institute（米拉-魁北克人工智能研究所）；Polytechnique Montréal（蒙特利尔理工学院）；Amazon Alexa（亚马逊Alexa）；University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）
University of Southern California（南加州大学）；University of Chicago（芝加哥大学）；Temple University（天普大学）；University of Birmingham（伯明翰大学）；Toyota Technological Institute at Chicago（芝加哥丰田技术研究院）；Penn State University（宾夕法尼亚州立大学）
Standard architectures used in instruction following often struggle on novel compositions of subgoals (e.g. navigating to landmarks or picking up objects) observed during training. We propose a modular architecture for following natural language instructions that describe sequences of diverse subgoals. In our approach, subgoal modules each carry out natural language instructions for a specific subgoal type. A sequence of modules to execute is chosen by learning to segment the instructions and predicting a subgoal type for each segment. When compared to standard, non-modular sequence-to-sequenc…
Robin3D: Improving 3D Large Language Model via Robust Instruction Tuning
University of Illinois Chicago（伊利诺伊大学芝加哥分校）；Zhejiang University（浙江大学）；University of Central Florida（中佛罗里达大学）；Illinois Institute of Technology（伊利诺伊理工大学）
TAL Education Group（好未来教育集团）；Guangdong Institute of Smart Education（广东智慧教育研究院）；Jinan University（暨南大学）
Online dialogic instructions are a set of pedagogical instructions used in real-world online educational contexts to motivate students, help understand learning materials, and build effective study habits. In spite of the popularity and advantages of online learning, the education technology and educational data mining communities still suffer from the lack of large-scale, high-quality, and well-annotated teaching instruction datasets to study computational approaches to automatically detect online dialogic instructions and further improve the online teaching effectiveness. Therefore, in this …
Generalization-Enhanced Code Vulnerability Detection via Multi-Task Instruction Fine-Tuning
School of Cyber Science and Engineering, Huazhong University of Science and Technology（华中科技大学网络空间科学与技术学院）；College of Computer, National University of Defense Technology（国防科技大学计算机学院）；School of Computer Science and Technology, Huazhong University of Science and Technology（华中科技大学计算机科学与技术学院）
CLASS-IT: Conversational and Lecture-Aligned Small-Scale Instruction Tuning for BabyLMs
CoLing Lab, Department of Philology, Literature and Linguistics, University of Pisa（皮尔切大学哲学、文学与语言学系语言实验室）；Department of Computer Science, University of Pisa（皮尔切大学计算机科学系）
Unseen from Seen: Rewriting Observation-Instruction Using Foundation Models for Augmenting Vision-Language Navigation
Shenzhen Campus of Sun Yat-sen University（中山大学深圳校区）；Shanghai Jiao Tong University（上海交通大学）；The University of Hong Kong（香港大学）；Hunan Artificial Intelligence and Robotics Institute Company Ltd.（湖南人工智能与机器人研究院有限公司）；Huawei Noah’s Ark Lab（华为诺亚实验室）；Peng Cheng Laboratory（鹏城实验室）
Towards Compositional Generalization of LLMs via Skill Taxonomy Guided Data Synthesis
State Key Laboratory of Complex & Critical Software Environment, Beihang University（复杂与关键软件环境国家重点实验室，北京航空航天大学）；Beijing Academy of Artificial Intelligence（北京人工智能研究院）；Beijing Institute of Technology（北京理工大学）；Institute of Computing Technology, CAS（中国科学院计算技术研究所）
Department of Computer Science, University of Southern California（南加州大学计算机科学系）；Facebook Artificial Intelligence Research（Facebook人工智能研究院）
The recursive Newton-Euler Algorithm (RNEA) is a popular technique for computing the dynamics of robots. RNEA can be framed as a differentiable computational graph, enabling the dynamics parameters of the robot to be learned from data via modern auto-differentiation toolboxes. However, the dynamics parameters learned in this manner can be physically implausible. In this work, we incorporate physical constraints in the learning by adding structure to the learned parameters. This results in a framework that can learn physically plausible dynamics via gradient descent, improving the training spee…
JsonTuning: Towards Generalizable, Robust, and Controllable Instruction Tuning
The Chinese University of Hong Kong（香港中文大学）；Singapore University of Technology and Design（新加坡科技设计大学）；DAMO Academy, Alibaba Group（阿里云达摩院）；Nanyang Technological University（南洋理工大学）
What Makes for Good Visual Instructions? Synthesizing Complex Visual Reasoning Instructions for Visual Instruction Tuning
Gaoling School of Artificial Intelligence, Renmin University of China（中国人民大学高瓴人工智能学院）；School of Information, Renmin University of China（中国人民大学信息学院）；Meituan Group（美团集团）
Allen Institute for AI（艾伦人工智能研究所）；Univ. of Washington（华盛顿大学）；Arizona State Univ.（亚利桑那州立大学）；Sharif Univ. of Tech.（谢里夫理工大学）；Tehran Polytechnic（德黑兰理工学院）；PSG College of Tech.（PSG理工学院）；IIT Kharagpur（印度理工学院克勒格布尔分校）；Univ. of Amsterdam（阿姆斯特丹大学）；UC Berkeley（加州大学伯克利分校）；Columbia Univ.（哥伦比亚大学）；Factored AI；Govt. Polytechnic Rajkot（拉杰科特政府理工学院）；Microsoft Research（微软研究院）；Stanford Univ.（斯坦福大学）；Zycus Infotech（Zycus信息技术公司）；Univ. of Massachusetts Amherst（马萨诸塞大学阿默斯特分校）；National Inst. of Tech. Karnataka（卡纳塔克国家理工学院）；TCS Research（塔塔咨询服务研究院）；IIT Madras（印度理工学院马德拉斯分校）；National Univ. of Singapore（新加坡国立大学）；Johns Hopkins Univ.（约翰斯·霍普金斯大学）
How well can NLP models generalize to a variety of unseen tasks when provided with task instructions? To address this question, we first introduce Super-NaturalInstructions, a benchmark of 1,616 diverse NLP tasks and their expert-written instructions. Our collection covers 76 distinct task types, including but not limited to classification, extraction, infilling, sequence tagging, text rewriting, and text composition. This large and diverse collection of tasks enables rigorous benchmarking of cross-task generalization under instructions -- training models to follow instructions on a subset of …
