Multi-turn interactions between large language models (LLMs) and users naturally include implicit feedback signals. If an LLM responds in an unexpected way to an instruction, the user is likely to signal it by rephrasing the request, expressing frustration, or pivoting to an alternative task. Such signals are task-independent and occupy a relatively constrained subspace of language, allowing the LLM to identify them even if it fails on the actual task. We introduce ReSpect, a method to learn from such signals in past interactions via retrospection without additional annotations. We deploy ReSpect in a new multimodal interaction scenario, where humans instruct a multimodal LLM to solve an abstract reasoning task with a combinatorial solution space. Through thousands of interactions with humans, we show how ReSpect gradually improves task completion rate from 31% to 82%, all without any external annotation.

We propose KnotGym, an interactive environment for complex, spatial reasoning and manipulation. KnotGym includes goal-oriented rope manipulation tasks with varying levels of complexity, all requiring acting from pure image observations.Tasks are defined along a clear and quantifiable axis of complexity based on the number of knot crossings, creating a natural generalization test.KnotGym has a simple observation space, allowing for scalable development, yet it highlights core challenges in integrating acute perception, spatial reasoning, and grounded manipulation.We evaluate methods of different classes, including model-based RL, model-predictive control, and chain-of-thought reasoning, and illustrate the challenges KnotGym presents.

Multispectral pedestrian detection has been shown to be effective in improving performance in complex illumination scenarios. However, prevalent double-stream networks in multispectral detection employ two separate feature extraction branches for multi-modal data, leading to nearly double the inference time compared to single-stream networks utilizing only one feature extraction branch. This increased inference time has hindered the widespread employment of multispectral pedestrian detection in embedded devices for autonomous systems. To efficiently compress multispectral object detection networks, we propose a novel distillation method, the Adaptive Modal Fusion Distillation (AMFD) framework. Unlike traditional distillation methods, the AMFD framework fully leverages the original modal features from the teacher network, thereby significantly enhancing the performance of the student network. Specifically, a Modal Extraction Alignment (MEA) module is utilized to derive learning weights for student networks, integrating focal and global attention mechanisms. This methodology enables the student network to acquire optimal fusion strategies independent from that of teacher network without necessitating an additional feature fusion module. Furthermore, we present the SMOD dataset, a well-aligned challenging multispectral dataset for detection. Extensive experiments on the challenging KAIST, LLVIP, SUNRGB-D and SMOD datasets are conducted to validate the effectiveness of AMFD. The results demonstrate that our method outperforms existing state-of-the-art methods in both reducing log-average Miss Rate and improving mean Average Precision. The code is available athttps://github.com/bigD233/AMFD.git.

Reward Machines provide an automaton-inspired structure for specifying instructions, safety constraints, and other temporally extended reward-worthy behaviour. By exposing the underlying structure of a reward function, they enable the decomposition of an RL task, leading to impressive gains in sample efficiency. Although Reward Machines and similar formal specifications have a rich history of application towards sequential decision-making problems, prior frameworks have traditionally ignored ambiguity and uncertainty when interpreting the domain-specific vocabulary forming the building blocks of the reward function. Such uncertainty critically arises in many real-world settings due to factors like partial observability or noisy sensors. In this work, we explore the use of Reward Machines for Deep RL in noisy and uncertain environments. We characterize this problem as a POMDP and propose a suite of RL algorithms that exploit task structure under uncertain interpretation of the domain-specific vocabulary. Through theory and experiments, we expose pitfalls in naive approaches to this problem while simultaneously demonstrating how task structure can be successfully leveraged under noisy interpretations of the vocabulary.

Aerospace components (ACs) are an important part of the aerospace equipment. Because of process specialty, AC production scheduling has always been seen a challenging work. In this paper, an improved dual-resource-constrained flexible job shop scheduling model is constructed to formulate the problem. The characteristics of AC manufacturing process are fully considered by analyzing the relationship among processes, machines, and workers. Meanwhile, the effect of machine type, worker skill level and worker labor efficiency are also incorporated into the model. To solve the model, we design an evolutionary strategy combined with TOPSIS and used Metropolis guidelines and perturbation operators to ameliorate the search process. A two-stage search-enhanced multi-objective evolutionary algorithm is proposed, which aims to minimize the completion time, production cost and worker load imbalance. Finally, experiments are conducted based on real AC production data. Experimental results verify the proposed algorithm is effective and competitive, which provides certain application value for relevant enterprises.

Benefiting from non-invasive sensing tech- nologies, heartbeat detection from ballistocardiogram (BCG) signals is of great significance for home-care applications, such as risk prediction of cardiovascular disease (CVD) and sleep staging, etc. In this paper, we propose an effective deep learning model for automatic heartbeat detection from BCG signals based on UNet and bidirectional long short-term memory (Bi-LSTM). The developed deep learning model provides an effective solution to the existing challenges in BCG-aided heartbeat detection, especially for BCG in low signal-to-noise ratio, in which the waveforms in BCG signals are irregular due to measured postures, rhythm and artifact motion. For validations, performance of the proposed detection is evaluated by BCG recordings from 43 subjects with different measured postures and heart rate ranges. The accuracy of the detected heartbeat intervals measured in different postures and signal qualities, in comparison with the R-R interval of ECG, is promising in terms of mean absolute error and mean relative error, respectively, which is superior to the state-of-the-art methods. Numerical results demonstrate that the proposed UNet-BiLSTM model performs robust to noise and perturbations (e.g. respiratory effort and artifact motion) in BCG signals, and provides a reliable solution to long term heart rate monitoring.

Even though earliest-deadline-first (EDF) is optimal in terms of uniprocessor schedulability, it is co-NP-hard to precisely verify uniprocessor schedulability for constrained-deadline task sets. The most efficient way to solve this problem in polynomial time is via a partially linear approximation of the demand bound function. Such approximation leads to a simple uniprocessor schedulability testing with speedup factor ρ. Such a result further leads to Deadline-Monotonic Partitioned-EDF on multi-processors with speedup factor of 1 + ρ − 1/m (where m is the number of processors). The current state of the art results indicate that ρ is within the range [1.5,14/9]. Especially, it has been a conjecture that ρ = 1.5.This paper improves the range of ρ to (1.5026,1.5380). The improved lower bound disproves the conjecture of lower bound 1.5. A novel technique is to construct an auxiliary function that is larger than the approximate demand bound function but keeps the supremum ρ unchanged. It solves the dilemma that beating the lower bound 1.5 requires extremely large task sets, while the large size makes it difficult to check the schedulability. This technique not only enables us to disprove 1.5 by a task set of only eight tasks, but also sheds light on future work in transferring/downsizing task sets and deriving utilization bound based tests for various workload abstraction models, such as DAG tasks.