This article surveys research works in the quickly advancing field of instruction tuning (IT), a crucial technique to enhance the capabilities and controllability of large language models (LLMs). Instruction tuning refers to the process of further training LLMs on a dataset consisting of (instruction, output) pairs in a supervised fashion, which bridges the gap between the next-word prediction objective of LLMs and the users’ objective of having LLMs adhere to human instructions. In this work, we make a systematic review of the literature, including the general methodology of IT, the construction of IT datasets, the training of IT models, and applications to different modalities, domains and application, along with analysis of aspects that influence the outcome of IT (e.g., generation of instruction outputs, size of the instruction dataset). We also review the potential pitfalls of IT along with criticism against it, along with efforts pointing out current deficiencies of existing strategies and suggest some avenues for fruitful research.

In this paper, we review recent emerging theoretical and technological advances of artificial intelligence (AI) in the big data settings. We conclude that integrating data-driven machine learning with human knowledge (common priors or implicit intuitions) can effectively lead to explainable, robust, and general AI, as follows: from shallow computation to deep neural reasoning; from merely data-driven model to data-driven with structured logic rules models; from task-oriented (domain-specific) intelligence (adherence to explicit instructions) to artificial general intelligence in a general context (the capability to learn from experience). Motivated by such endeavors, the next generation of AI, namely AI 2.0, is positioned to reinvent computing itself, to transform big data into structured knowledge, and to enable better decision-making for our society.

The segmentation of pancreas is important for medical image analysis, yet it faces great challenges of class imbalance, background distractions, and non-rigid geometrical features. To address these difficulties, we introduce a deep Q network (DQN) driven approach with deformable U-Net to accurately segment the pancreas by explicitly interacting with contextual information and extract anisotropic features from pancreas. The DQN-based model learns a context-adaptive localization policy to produce a visually tightened and precise localization bounding box of the pancreas. Furthermore, deformable U-Net captures geometry-aware information of pancreas by learning geometrically deformable filters for feature extraction. The experiments on NIH dataset validate the effectiveness of the proposed framework in pancreas segmentation.

Background. Solving the class-imbalance problem of within-project software defect prediction (SDP) is an important research topic. Although some class-imbalance learning methods have been presented, there exists room for improvement. For cross-project SDP, we found that the class-imbalanced source usually leads to misclassification of defective instances. However, only one work has paid attention to this cross-project class-imbalance problem. Objective. We aim to provide effective solutions for both within-project and cross-project class-imbalance problems. Method. Subclass discriminant analysis (SDA), an effective feature learning method, is introduced to solve the problems. It can learn features with more powerful classification ability from original metrics. For within-project prediction, we improve SDA for achieving balanced subclasses and propose the improved SDA (ISDA) approach. For cross-project prediction, we employ the semi-supervised transfer component analysis (SSTCA) method to make the distributions of source and target data consistent, and propose the SSTCA+ISDA prediction approach. Results . Extensive experiments on four widely used datasets indicate that: 1) ISDA-based solution performs better than other state-of-the-art methods for within-project class-imbalance problem; 2) SSTCA+ISDA proposed for cross-project class-imbalance problem significantly outperforms related methods. Conclusion . Within-project and cross-project class-imbalance problems greatly affect prediction performance, and we provide a unified and effective prediction framework for both problems.

As a challenging problem, few-shot class-incremental learning (FSCIL) continually learns a sequence of tasks, confronting the dilemma between slow forgetting of old knowledge and fast adaptation to new knowledge. In this paper, we concentrate on this “slow versus fast” (SvF) dilemma to determine which knowledge components to be updated in a slow fashion or a fast fashion, and thereby balance old-knowledge preservation and new-knowledge adaptation. We propose a multi-grained SvF learning strategy to cope with the SvF dilemma from two different grains: intra-space (within the same feature space) and inter-space (between two different feature spaces). The proposed strategy designs a novel frequency-aware regularization to boost the intra-space SvF capability, and meanwhile develops a new feature space composition operation to enhance the inter-space SvF learning performance. With the multi-grained SvF learning strategy, our method outperforms the state-of-the-art approaches by a large margin.

With the memory-resource-limited constraints, class-incremental learning (CIL) usually suffers from the “catastrophic forgetting” problem when updating the joint classification model on the arrival of newly added classes. To cope with the forgetting problem, many CIL methods transfer the knowledge of old classes by preserving some exemplar samples into the size-constrained memory buffer. To utilize the memory buffer more efficiently, we propose to keep more auxiliary low-fidelity exemplar samples, rather than the original real-high-fidelity exemplar samples. Such a memory-efficient exemplar preserving scheme makes the old-class knowledge transfer more effective. However, the low-fidelity exemplar samples are often distributed in a different domain away from that of the original exemplar samples, that is, a domain shift. To alleviate this problem, we propose a duplet learning scheme that seeks to construct domain-compatible feature extractors and classifiers, which greatly narrows down the above domain gap. As a result, these low-fidelity auxiliary exemplar samples have the ability to moderately replace the original exemplar samples with a lower memory cost. In addition, we present a robust classifier adaptation scheme, which further refines the biased classifier (learned with the samples containing distillation label knowledge about old classes) with the help of the samples of pure true class labels. Experimental results demonstrate the effectiveness of this work against the state-of-the-art approaches. We will release the code, baselines, and training statistics for all models to facilitate future research.