Zero-Shot Generalization during Instruction Tuning: Insights from Similarity and Granularity

Understanding alignment techniques begins with comprehending zero-shot generalization brought by instruction tuning, but little of the mechanism has been understood. Existing work has largely been confined to the task level, without considering that tasks are artificially defined and, to LLMs, merely consist of tokens and representations. This line of research has been limited to examining transfer between tasks from a task-pair perspective, with few studies focusing on understanding zero-shot generalization from the perspective of the data itself. To bridge this gap, we first demonstrate through multiple metrics that zero-shot generalization during instruction tuning happens very early. Next, we investigate the facilitation of zero-shot generalization from both data similarity and granularity perspectives, confirming that encountering highly similar and fine-grained training data earlier during instruction tuning, without the constraints of defined "tasks", enables better generalization. Finally, we propose a more grounded training data arrangement method, Test-centric Multi-turn Arrangement, and show its effectiveness in promoting continual learning and further loss reduction. For the first time, we show that zero-shot generalization during instruction tuning is a form of similarity-based generalization between training and test data at the instance level. We hope our analysis will advance the understanding of zero-shot generalization during instruction tuning and contribute to the development of more aligned LLMs. Our code is released at https://github.com/HBX-hbx/dynamics_of_zero-shot_generalization.

Autonomous Path Planning for Intercostal Robotic Ultrasound Imaging Using Reinforcement Learning

Ultrasound (US) has been widely used in daily clinical practice for screening internal organs and guiding interventions. However, due to the acoustic shadow cast by the subcutaneous rib cage, the US examination for thoracic application is still challenging. To fully cover and reconstruct the region of interest in US for diagnosis, an intercostal scanning path is necessary. To tackle this challenge, we present a reinforcement learning (RL) approach for planning scanning paths between ribs to monitor changes in lesions on internal organs, such as the liver and heart, which are covered by rib cages. Structured anatomical information of the human skeleton is crucial for planning these intercostal paths. To obtain such anatomical insight, an RL agent is trained in a virtual environment constructed using computational tomography (CT) templates with randomly initialized tumors of various shapes and locations. In addition, task-specific state representation and reward functions are introduced to ensure the convergence of the training process while minimizing the effects of acoustic attenuation and shadows during scanning. To validate the effectiveness of the proposed approach, experiments have been carried out on unseen CTs with randomly defined single or multiple scanning targets. The results demonstrate the efficiency of the proposed RL framework in planning non-shadowed US scanning trajectories in areas with limited acoustic access.

LanePtrNet: Revisiting Lane Detection as Point Voting and Grouping on Curves

Lane detection plays a critical role in the field of autonomous driving. Prevailing methods generally adopt basic concepts (anchors, key points, etc.) from object detection and segmentation tasks, while these approaches require manual adjustments for curved objects, involve exhaustive searches on predefined anchors, require complex post-processing steps, and may lack flexibility when applied to real-world scenarios.In this paper, we propose a novel approach, LanePtrNet, which treats lane detection as a process of point voting and grouping on ordered sets: Our method takes backbone features as input and predicts a curve-aware centerness, which represents each lane as a point and assigns the most probable center point to it. A novel point sampling method is proposed to generate a set of candidate points based on the votes received. By leveraging features from local neighborhoods, and cross-instance attention score, we design a grouping module that further performs lane-wise clustering between neighboring and seeding points. Furthermore, our method can accommodate a point-based framework, (PointNet++ series, etc.) as an alternative to the backbone. This flexibility enables effortless extension to 3D lane detection tasks. We conduct comprehensive experiments to validate the effectiveness of our proposed approach, demonstrating its superior performance.

Reconstruction-Based Anomaly Localization via Knowledge-Informed Self-Training

Anomaly localization, which involves localizing anomalous regions within images, is a significant industrial task. Reconstruction-based methods are widely adopted for anomaly localization because of their low complexity and high interpretability. Most existing reconstruction-based methods only use normal samples to construct model. If anomalous samples are appropriately utilized in the process of anomaly localization, the localization performance can be improved. However, usually only weakly labeled anomalous samples are available, which limits the improvement. In many cases, we can obtain some knowledge of anomalies summarized by domain experts. Taking advantage of such knowledge can help us better utilize the anomalous samples and thus further improve the localization performance. In this paper, we propose a novel reconstruction-based method named knowledge-informed self-training (KIST) which integrates knowledge into reconstruction model through self-training. Specifically, KIST utilizes weakly labeled anomalous samples in addition to the normal ones and exploits knowledge to yield pixel-level pseudo-labels of the anomalous samples. Based on the pseudo labels, a novel loss which promotes the reconstruction of normal pixels while suppressing the reconstruction of anomalous pixels is used. We conduct experiments on different datasets and demonstrate the advantages of KIST over the existing reconstruction-based methods.

Tell Me More! Towards Implicit User Intention Understanding of Language Model Driven Agents

Current language model-driven agents often lack mechanisms for effective user participation, which is crucial given the vagueness commonly found in user instructions. Although adept at devising strategies and performing tasks, these agents struggle with seeking clarification and grasping precise user intentions. To bridge this gap, we introduce Intention-in-Interaction (IN3), a novel benchmark designed to inspect users' implicit intentions through explicit queries. Next, we propose the incorporation of model experts as the upstream in agent designs to enhance user-agent interaction. Employing IN3, we empirically train Mistral-Interact, a powerful model that proactively assesses task vagueness, inquires user intentions, and refines them into actionable goals before starting downstream agent task execution. Integrating it into the XAgent framework, we comprehensively evaluate the enhanced agent system regarding user instruction understanding and execution, revealing that our approach notably excels at identifying vague user tasks, recovering and summarizing critical missing information, setting precise and necessary agent execution goals, and minimizing redundant tool usage, thus boosting overall efficiency. All the data and codes are released.

Investigate-Consolidate-Exploit: A General Strategy for Inter-Task Agent Self-Evolution

This paper introduces Investigate-Consolidate-Exploit (ICE), a novel strategy for enhancing the adaptability and flexibility of AI agents through inter-task self-evolution. Unlike existing methods focused on intra-task learning, ICE promotes the transfer of knowledge between tasks for genuine self-evolution, similar to human experience learning. The strategy dynamically investigates planning and execution trajectories, consolidates them into simplified workflows and pipelines, and exploits them for improved task execution. Our experiments on the XAgent framework demonstrate ICE's effectiveness, reducing API calls by as much as 80% and significantly decreasing the demand for the model's capability. Specifically, when combined with GPT-3.5, ICE's performance matches that of raw GPT-4 across various agent tasks. We argue that this self-evolution approach represents a paradigm shift in agent design, contributing to a more robust AI community and ecosystem, and moving a step closer to full autonomy.

Triple Simplex Matrix Completion for Expense Forecasting

Forecasting project expenses is a crucial step for businesses to avoid budget overruns and project failures. Traditionally, this has been done by financial analysts or data science techniques such as time-series analysis. However, these approaches can be uncertain and produce results that differ from the planned budget, especially at the start of a project with limited data points. This paper proposes a constrained non-negative matrix completion model that predicts expenses by learning the likelihood of the project correlating with certain expense patterns in the latent space. The model is constrained on three probability simplexes, two of which are on the factor matrices and the third on the missing entries. Additionally, the predicted expense values are guaranteed to meet the budget constraint without the need of post-processing. An inexact alternating optimization algorithm is developed to solve the associated optimization problem and is proven to converge to a stationary point. Results from two real datasets demonstrate the effectiveness of the proposed method in comparison to state-of-the-art algorithms.

Toolink: Linking Toolkit Creation and Using through Chain-of-Solving on Open-Source Model

Large Language Models (LLMs) have demonstrated remarkable progress in utilizing tools, but their closed-source nature and high inference costs pose limitations on their adaptability, necessitating a valid method that leverages smaller, open-sourced models. In this paper, we introduce Toolink, a comprehensive framework that performs task-solving by first creating a toolkit and then integrating the planning and calling of tools through a chain-of-solving (CoS) approach. We first validate the efficacy of Toolink in harnessing the model's creativity and CoS ability on ChatGPT. Subsequently, we curate CoS-GPT, a chain-of-solving dataset designed for tool-using, and finetune the LLaMA-7B model. It results in LLaMA-CoS, a powerful open-source model with advanced tool-planning and tool-calling capabilities. Evaluation on diverse tasks from BIG-bench demonstrates its CoS ability matches that of ChatGPT while its performance surpasses the chain-of-thought approach. Further studies highlight the generalization of LLaMA-CoS to unseen tasks and showcase its capability in using toolkits not explicitly tailored for the target task, affirming its robustness in real-world scenarios. All codes and data are released.

"Merge Conflicts!" Exploring the Impacts of External Distractors to Parametric Knowledge Graphs

Large language models (LLMs) acquire extensive knowledge during pre-training, known as their parametric knowledge. However, in order to remain up-to-date and align with human instructions, LLMs inevitably require external knowledge during their interactions with users. This raises a crucial question: How will LLMs respond when external knowledge interferes with their parametric knowledge? To investigate this question, we propose a framework that systematically elicits LLM parametric knowledge and introduces external knowledge. Specifically, we uncover the impacts by constructing a parametric knowledge graph to reveal the different knowledge structures of LLMs, and introduce external knowledge through distractors of varying degrees, methods, positions, and formats. Our experiments on both black-box and open-source models demonstrate that LLMs tend to produce responses that deviate from their parametric knowledge, particularly when they encounter direct conflicts or confounding changes of information within detailed contexts. We also find that while LLMs are sensitive to the veracity of external knowledge, they can still be distracted by unrelated information. These findings highlight the risk of hallucination when integrating external knowledge, even indirectly, during interactions with current LLMs. All the data and results are publicly available.

Deep Reinforcement Learning in Surgical Robotics: Enhancing the Automation Level

Surgical robotics is a rapidly evolving field that is transforming the landscape of surgeries. Surgical robots have been shown to enhance precision, minimize invasiveness, and alleviate surgeon fatigue. One promising area of research in surgical robotics is the use of reinforcement learning to enhance the automation level. Reinforcement learning is a type of machine learning that involves training an agent to make decisions based on rewards and punishments. This literature review aims to comprehensively analyze existing research on reinforcement learning in surgical robotics. The review identified various applications of reinforcement learning in surgical robotics, including pre-operative, intra-body, and percutaneous procedures, listed the typical studies, and compared their methodologies and results. The findings show that reinforcement learning has great potential to improve the autonomy of surgical robots. Reinforcement learning can teach robots to perform complex surgical tasks, such as suturing and tissue manipulation. It can also improve the accuracy and precision of surgical robots, making them more effective at performing surgeries.