Department of Computer Engineering, School of Computation, Information and Technology, Technical University of Munich
Abstract:In frame-based vision, object detection faces substantial performance degradation under challenging conditions due to the limited sensing capability of conventional cameras. Event cameras output sparse and asynchronous events, providing a potential solution to solve these problems. However, effectively fusing two heterogeneous modalities remains an open issue. In this work, we propose a novel hierarchical feature refinement network for event-frame fusion. The core concept is the design of the coarse-to-fine fusion module, denoted as the cross-modality adaptive feature refinement (CAFR) module. In the initial phase, the bidirectional cross-modality interaction (BCI) part facilitates information bridging from two distinct sources. Subsequently, the features are further refined by aligning the channel-level mean and variance in the two-fold adaptive feature refinement (TAFR) part. We conducted extensive experiments on two benchmarks: the low-resolution PKU-DDD17-Car dataset and the high-resolution DSEC dataset. Experimental results show that our method surpasses the state-of-the-art by an impressive margin of $\textbf{8.0}\%$ on the DSEC dataset. Besides, our method exhibits significantly better robustness (\textbf{69.5}\% versus \textbf{38.7}\%) when introducing 15 different corruption types to the frame images. The code can be found at the link (https://github.com/HuCaoFighting/FRN).
Abstract:3D point cloud segmentation has received significant interest for its growing applications. However, the generalization ability of models suffers in dynamic scenarios due to the distribution shift between test and training data. To promote robustness and adaptability across diverse scenarios, test-time adaptation (TTA) has recently been introduced. Nevertheless, most existing TTA methods are developed for images, and limited approaches applicable to point clouds ignore the inherent hierarchical geometric structures in point cloud streams, i.e., local (point-level), global (object-level), and temporal (frame-level) structures. In this paper, we delve into TTA in 3D point cloud segmentation and propose a novel Hierarchical Geometry Learning (HGL) framework. HGL comprises three complementary modules from local, global to temporal learning in a bottom-up manner.Technically, we first construct a local geometry learning module for pseudo-label generation. Next, we build prototypes from the global geometry perspective for pseudo-label fine-tuning. Furthermore, we introduce a temporal consistency regularization module to mitigate negative transfer. Extensive experiments on four datasets demonstrate the effectiveness and superiority of our HGL. Remarkably, on the SynLiDAR to SemanticKITTI task, HGL achieves an overall mIoU of 46.91\%, improving GIPSO by 3.0\% and significantly reducing the required adaptation time by 80\%. The code is available at https://github.com/tpzou/HGL.
Abstract:Learning from demonstrations faces challenges in generalizing beyond the training data and is fragile even to slight visual variations. To tackle this problem, we introduce Lan-o3dp, a language guided object centric diffusion policy that takes 3d representation of task relevant objects as conditional input and can be guided by cost function for safety constraints at inference time. Lan-o3dp enables strong generalization in various aspects, such as background changes, visual ambiguity and can avoid novel obstacles that are unseen during the demonstration process. Specifically, We first train a diffusion policy conditioned on point clouds of target objects and then harness a large language model to decompose the user instruction into task related units consisting of target objects and obstacles, which can be used as visual observation for the policy network or converted to a cost function, guiding the generation of trajectory towards collision free region at test time. Our proposed method shows training efficiency and higher success rates compared with the baselines in simulation experiments. In real world experiments, our method exhibits strong generalization performance towards unseen instances, cluttered scenes, scenes of multiple similar objects and demonstrates training free capability of obstacle avoidance.
Abstract:As transformer-based object detection models progress, their impact in critical sectors like autonomous vehicles and aviation is expected to grow. Soft errors causing bit flips during inference have significantly impacted DNN performance, altering predictions. Traditional range restriction solutions for CNNs fall short for transformers. This study introduces the Global Clipper and Global Hybrid Clipper, effective mitigation strategies specifically designed for transformer-based models. It significantly enhances their resilience to soft errors and reduces faulty inferences to ~ 0\%. We also detail extensive testing across over 64 scenarios involving two transformer models (DINO-DETR and Lite-DETR) and two CNN models (YOLOv3 and SSD) using three datasets, totalling approximately 3.3 million inferences, to assess model robustness comprehensively. Moreover, the paper explores unique aspects of attention blocks in transformers and their operational differences from CNNs.
Abstract:We introduce Situation Monitor, a novel zero-shot Out-of-Distribution (OOD) detection approach for transformer-based object detection models to enhance reliability in safety-critical machine learning applications such as autonomous driving. The Situation Monitor utilizes the Diversity-based Budding Ensemble Architecture (DBEA) and increases the OOD performance by integrating a diversity loss into the training process on top of the budding ensemble architecture, detecting Far-OOD samples and minimizing false positives on Near-OOD samples. Moreover, utilizing the resulting DBEA increases the model's OOD performance and improves the calibration of confidence scores, particularly concerning the intersection over union of the detected objects. The DBEA model achieves these advancements with a 14% reduction in trainable parameters compared to the vanilla model. This signifies a substantial improvement in efficiency without compromising the model's ability to detect OOD instances and calibrate the confidence scores accurately.
Abstract:Recently, Large Language Models (LLMs) have witnessed remarkable performance as zero-shot task planners for robotic manipulation tasks. However, the open-loop nature of previous works makes LLM-based planning error-prone and fragile. On the other hand, failure detection approaches for closed-loop planning are often limited by task-specific heuristics or following an unrealistic assumption that the prediction is trustworthy all the time. As a general-purpose reasoning machine, LLMs or Multimodal Large Language Models (MLLMs) are promising for detecting failures. However, However, the appropriateness of the aforementioned assumption diminishes due to the notorious hullucination problem. In this work, we attempt to mitigate these issues by introducing a framework for closed-loop LLM-based planning called KnowLoop, backed by an uncertainty-based MLLMs failure detector, which is agnostic to any used MLLMs or LLMs. Specifically, we evaluate three different ways for quantifying the uncertainty of MLLMs, namely token probability, entropy, and self-explained confidence as primary metrics based on three carefully designed representative prompting strategies. With a self-collected dataset including various manipulation tasks and an LLM-based robot system, our experiments demonstrate that token probability and entropy are more reflective compared to self-explained confidence. By setting an appropriate threshold to filter out uncertain predictions and seek human help actively, the accuracy of failure detection can be significantly enhanced. This improvement boosts the effectiveness of closed-loop planning and the overall success rate of tasks.
Abstract:Safe reinforcement learning (RL) is crucial for deploying RL agents in real-world applications, as it aims to maximize long-term rewards while satisfying safety constraints. However, safe RL often suffers from sample inefficiency, requiring extensive interactions with the environment to learn a safe policy. We propose Efficient Safe Policy Optimization (ESPO), a novel approach that enhances the efficiency of safe RL through sample manipulation. ESPO employs an optimization framework with three modes: maximizing rewards, minimizing costs, and balancing the trade-off between the two. By dynamically adjusting the sampling process based on the observed conflict between reward and safety gradients, ESPO theoretically guarantees convergence, optimization stability, and improved sample complexity bounds. Experiments on the Safety-MuJoCo and Omnisafe benchmarks demonstrate that ESPO significantly outperforms existing primal-based and primal-dual-based baselines in terms of reward maximization and constraint satisfaction. Moreover, ESPO achieves substantial gains in sample efficiency, requiring 25--29% fewer samples than baselines, and reduces training time by 21--38%.
Abstract:In numerous reinforcement learning (RL) problems involving safety-critical systems, a key challenge lies in balancing multiple objectives while simultaneously meeting all stringent safety constraints. To tackle this issue, we propose a primal-based framework that orchestrates policy optimization between multi-objective learning and constraint adherence. Our method employs a novel natural policy gradient manipulation method to optimize multiple RL objectives and overcome conflicting gradients between different tasks, since the simple weighted average gradient direction may not be beneficial for specific tasks' performance due to misaligned gradients of different task objectives. When there is a violation of a hard constraint, our algorithm steps in to rectify the policy to minimize this violation. We establish theoretical convergence and constraint violation guarantees in a tabular setting. Empirically, our proposed method also outperforms prior state-of-the-art methods on challenging safe multi-objective reinforcement learning tasks.
Abstract:Accurate and effective 3D object detection is critical for ensuring the driving safety of autonomous vehicles. Recently, state-of-the-art two-stage 3D object detectors have exhibited promising performance. However, these methods refine proposals individually, ignoring the rich contextual information in the object relationships between the neighbor proposals. In this study, we introduce an object relation module, consisting of a graph generator and a graph neural network (GNN), to learn the spatial information from certain patterns to improve 3D object detection. Specifically, we create an inter-object relationship graph based on proposals in a frame via the graph generator to connect each proposal with its neighbor proposals. Afterward, the GNN module extracts edge features from the generated graph and iteratively refines proposal features with the captured edge features. Ultimately, we leverage the refined features as input to the detection head to obtain detection results. Our approach improves upon the baseline PV-RCNN on the KITTI validation set for the car class across easy, moderate, and hard difficulty levels by 0.82%, 0.74%, and 0.58%, respectively. Additionally, our method outperforms the baseline by more than 1% under the moderate and hard levels BEV AP on the test server.
Abstract:Ensuring the safety of Reinforcement Learning (RL) is crucial for its deployment in real-world applications. Nevertheless, managing the trade-off between reward and safety during exploration presents a significant challenge. Improving reward performance through policy adjustments may adversely affect safety performance. In this study, we aim to address this conflicting relation by leveraging the theory of gradient manipulation. Initially, we analyze the conflict between reward and safety gradients. Subsequently, we tackle the balance between reward and safety optimization by proposing a soft switching policy optimization method, for which we provide convergence analysis. Based on our theoretical examination, we provide a safe RL framework to overcome the aforementioned challenge, and we develop a Safety-MuJoCo Benchmark to assess the performance of safe RL algorithms. Finally, we evaluate the effectiveness of our method on the Safety-MuJoCo Benchmark and a popular safe benchmark, Omnisafe. Experimental results demonstrate that our algorithms outperform several state-of-the-art baselines in terms of balancing reward and safety optimization.