Papers

The coexistence of heterogeneous service classes in 5G Enhanced Mobile Broadband (eMBB), Ultra-Reliable Low Latency Communication (URLLC), and Massive Machine-Type Communication (mMTC) poses major challenges for meeting diverse Quality-of-Service (QoS) requirements under limited spectrum and power resources. Existing radio access network (RAN) slicing schemes typically optimise isolated layers or objectives, lacking physical-layer realism, slot-level adaptability, and interpretable per-slice performance metrics. This paper presents a joint optimisation framework that integrates Dynamic Hybrid Resource Utilisation with MCS-Based Intelligent Layering, formulated as a mixed-integer linear program (MILP) that jointly allocates bandwidth, power, and modulation and coding scheme (MCS) indices per slice. The model incorporates finite blocklength effects, channel misreporting, and correlated fading to ensure realistic operation. Two modes are implemented: a Baseline Mode that ensures resource-efficient QoS feasibility, and an Ideal-Chaser Mode that minimises deviation from ideal per-slice rates. Simulation results show that the proposed approach achieves energy efficiencies above $10^7$~kb/J in Baseline Mode and sub-millisecond latency with near-ideal throughput in Ideal-Chaser Mode, outperforming recent optimisation and learning-based methods in delay, fairness, and reliability. The framework provides a unified, interpretable, and computationally tractable solution for dynamic cross-layer resource management in 5G and beyond networks.

This paper presents an $\mathcal{H}_\infty$ model matching control-based approach to the problem of power hardware-in-the-loop (PHIL) interfacing. The objective is to interconnect a grid simulation and a physical device via an interface in a way that is stable and accurate. Conventional approaches include the ideal transformer method (ITM) and its impedance-based variants, which trade accuracy for stability, as well as some $\mathcal{H}_\infty$ control-based approaches, which do not make use of all the available information in their optimization for accuracy. Designing for transparency, as opposed to accuracy as existing approaches do, would achieve both accuracy and stability, while making use of all the dynamical information present in the idealized interconnection of the grid and device. The approach proposed in this paper employs model matching to formulate the PHIL problem as an $\mathcal{H}_\infty$ control problem using transparency as the explicit frequency-domain control objective. The approach is experimentally validated in a real-time resistive-load PHIL setup, and is found to achieve accuracy levels that are comparable or superior to those of an ITM-based interface.

The recent advancements in autonomous and remote operation of maritime vessels necessitates the development of robust and reliable communication systems to support high-bandwidth applications such as real-time monitoring, navigation, and control. Existing communication channel models, including Rayleigh and Rician fading, are inadequate to accurately describe the dynamic and complex nature of maritime communication, particularly for high-speed vessels in coastal environments. This paper proposes an extension to the Multi-Cluster Fluctuating Two-Ray Fading (MFTR) model that also accounts for key phenomena such as large-scale fading, time-varying parameters and Doppler shifts. The extended MFTR model integrates Stochastic Differential Equations (SDEs) to capture the time-varying characteristics of the channel, such as phase shifts and delays, while considering physical factors like delay-induced power loss and path loss. The accuracy of the proposed model is assessed in simulation.

This paper presents the analytical design of a new wide tuning range and low-noise millimeter-wave voltage control oscillators (VCO) for 5G technology. The small signal model analysis and phase noise of the VCOs will be presented to evaluate the start-up oscillation condition, oscillation frequency, and phase noise affecting factors. Theoretical analysis and simulation results show the outperformance of the proposed cascode cross-couple LC VCO topology compared to the conventional cross-coupled LC VCO in terms of frequency tuning range, VCO gain and phase noise level.

Text-to-music generation technology is progressing rapidly, creating new opportunities for musical composition and editing. However, existing music editing methods often fail to preserve the source music's temporal structure, including melody and rhythm, when altering particular attributes like instrument, genre, and mood. To address this challenge, this paper conducts an in-depth probing analysis on attention maps within AudioLDM 2, a diffusion-based model commonly used as the backbone for existing music editing methods. We reveal a key finding: cross-attention maps encompass details regarding distinct musical characteristics, and interventions on these maps frequently result in ineffective modifications. In contrast, self-attention maps are essential for preserving the temporal structure of the source music during its conversion into the target music. Building upon this understanding, we present Melodia, a training-free technique that selectively manipulates self-attention maps in particular layers during the denoising process and leverages an attention repository to store source music information, achieving accurate modification of musical characteristics while preserving the original structure without requiring textual descriptions of the source music. Additionally, we propose two novel metrics to better evaluate music editing methods. Both objective and subjective experiments demonstrate that our approach achieves superior results in terms of textual adherence and structural integrity across various datasets. This research enhances comprehension of internal mechanisms within music generation models and provides improved control for music creation.

Data-driven health monitoring of power converters remains limited by poor generalization to unseen operating conditions. This work addresses this out-of-distribution (OOD) challenge by building a domain-specific time-series foundation model (PE-TSFM) that learns representations directly from large-scale unlabeled converter data. Unlike generic TSFMs trained on broad time-series datasets, the proposed PE-TSFM is pre-trained entirely on domain data, enabling it to learn the physical relationships unique to power electronics. To further tailor the model to this domain, we introduce a dual-attention mechanism that captures both temporal patterns and inter-channel dependencies. While generic TSFMs primarily model temporal dependencies, the added channel attention captures inter-sensor physical relationships essential for converter degradation analysis. A dataset containing 141 million unlabeled timestamps from an operating power converter is used for pre-training. Experiments show that PE-TSFM achieves 92% accuracy under unseen operating conditions. In contrast, generic TSFMs achieve around 60% and conventional time-series models achieve around 40% accuracy. This result confirms the strong OOD generalization of the proposed PE-TSFM. Ablation studies further verify that the introduced channel attention mechanism significantly improves model performance. In addition, we conduct detailed studies on model scalability, hyperparameter sensitivity, and interpretability to provide a comprehensive understanding of the proposed approach.

The emergence of 6G networks enables simultaneous near-field and far-field communications through extremely large antenna arrays and high carrier frequencies. While these regimes enhance spatial multiplexing and link capacity, their coexistence poses new challenges in ensuring quality-of-service (QoS) guarantees for delay-sensitive applications. This paper presents an effective capacity (EC) analysis framework that jointly models near- and far-field communication regimes under distance estimation uncertainty. The user location is modeled as a random variable spanning both propagation regions, and tractable closed-form expression for the EC is derived to quantify delay performance. Numerical results illustrate the impact of estimation variance, QoS exponent, far-field boundary and near-field boundary (Fraunhofer distance) on EC performance.

Standard model-based control design deteriorates when the system dynamics change during operation. To overcome this challenge, online and adaptive methods have been proposed in the literature. In this work, we consider the class of discrete-time linear systems with unknown time-varying parameters. We propose a simple, modular, and computationally tractable approach by combining two classical and well-known building blocks from estimation and control: the least mean square filter and the certainty-equivalent linear quadratic regulator. Despite both building blocks being simple and off-the-shelf, our analysis shows that they can be seamlessly combined to a powerful pipeline with stability guarantees. Namely, finite-gain $\ell^2$-stability of the closed-loop interconnection of the unknown system, the parameter estimator, and the controller is proven, despite the presence of unknown disturbances and time-varying parametric uncertainties. Real-world applicability of the proposed algorithm is showcased by simulations carried out on a nonlinear planar quadrotor.

6G networks are expected to integrate low Earth orbit satellites to ensure global connectivity by extending coverage to underserved and remote regions. However, the deployment of dense mega-constellations introduces severe interference among satellites operating over shared frequency bands. This is, in part, due to the limited flexibility of conventional frequency division duplex (FDD) systems, where fixed bands for downlink (DL) and uplink (UL) transmissions are employed. In this work, we propose dynamic re-assignment of FDD bands for improved interference management in dense deployments and evaluate the performance gain of this approach. To this end, we formulate a joint optimization problem that incorporates dynamic band assignment, user scheduling, and power allocation in both directions. This non-convex mixed integer problem is solved using a combination of equivalence transforms, alternating optimization, and state-of-the-art industrial-grade mixed integer solvers. Numerical results demonstrate that the proposed approach of dynamic FDD band assignment significantly enhances system performance over conventional FDD, achieving up to 94% improvement in throughput in dense deployments.

The change of electric power generation - from synchronous generator (SG) to converter - is generally regarded as the second revolution of power system. Different from rotor swing of SG in traditional grids mainly described by the swing equation (SE), the converter dynamics plays an indispensable role in modern renewable dominated power systems (RDPS). The high complexity of the RDPS, including spatial large-scale, nonlinearity, multi-time-scale, and even sequential switching, prevents us from fully understanding its dynamics and assessing its transient stability under large disturbance. Here, a variety of transient switching mechanism models of renewable devices relying on wind or solar energies under low-voltage ride-through are established and unified, which can be perfectly described by a generalized swing equation (GSE) under parameter changes for switching dynamics. The GSE focusing on the dominant phase-locking loop dynamics is similar to the SE. Mainly relying on the mechanical equivalence and the energy conservative principle, a substantially improved equal-area criterion method is proposed. Based on this method, even for large-scale renewable fields, the calculation errors for the critical clearing time are only about 1%. This elegant nonlinear-dynamics-based approach establishes a unified theory including modelling and analysis for the RDPS transient dynamics.

This paper presents an optimal power splitting and beamforming design for co-located simultaneous wireless information and power transfer (SWIPT) users in Dynamic Metasurface Antenna (DMA)-aided multiuser multiple-input single-output (MISO) systems. The objective is to minimize transmit power while meeting users signal-to-interference-plus-noise ratio (SINR) and energy harvesting (EH) requirements. The problem is solved via an alternating optimization framework based on semidefinite programming (SDP), where metasurface tunability follows Lorentzian-constrained holography (LCH). In contrast to traditional beamforming architectures, DMA-assisted architectures reduce the need for RF chains and phase shifters but require optimization under the Lorentzian constraint limiting the amplitude and phase optimizations. Hence, the proposed method integrates several LCH schemes, including the recently proposed adaptive-radius LCH (ARLCH), and evaluates nonlinear EH models and circuit noise effects. Simulation results show that the proposed design significantly reduces transmit power compared with baseline methods, highlighting the efficiency of ARLCH and optimal power splitting in DMA-assisted SWIPT systems.

This paper proposes a three-stage uplink channel estimation protocol for reconfigurable intelligent surface (RIS)-aided multi-user (MU) millimeter-wave (mmWave) multiple-input single-output (MISO) systems, where both the base station (BS) and the RIS are equipped with uniform planar arrays (UPAs). The proposed approach explicitly accounts for the mutual coupling (MC) effect, modeled via scattering parameter multiport network theory. In Stage~I, a dimension-reduced subspace-based method is proposed to estimate the common angle of arrival (AoA) at the BS using the received signals across all users. In Stage~II, MC-aware cascaded channel estimation is performed for a typical user. The equivalent measurement vectors for each cascaded path are extracted and the reference column is reconstructed using a compressed sensing (CS)-based approach. By leveraging the structure of the cascaded channel, the reference column is rearranged to estimate the AoA at the RIS, thereby reducing the computational complexity associated with estimating other columns. Additionally, the common angle of departure (AoD) at the RIS is also obtained in this stage, which significantly reduces the pilot overhead for estimating the cascaded channels of other users in Stage~III. The RIS phase shift training matrix is designed to optimize performance in the presence of MC and outperforms random phase scheme. Simulation results validate that the proposed method yields better performance than the MC-unaware and existing approaches in terms of estimation accuracy and pilot efficiency.

In this paper, we propose to employ a modular-based movable extremely large-scale array (XL-array) at Alice for enhancing covert communication performance. Compared with existing work that mostly considered either far-field or near-field covert communications, we consider in this paper a more general and practical mixed-field scenario, where multiple Bobs are located in either the near-field or far-field of Alice, in the presence of multiple near-field Willies. Specifically, we first consider a two-Bob-one-Willie system and show that conventional fixed-position XL-arrays suffer degraded sum-rate performance due to the energy-spread effect in mixed-field systems, which, however, can be greatly improved by subarray movement. On the other hand, for transmission covertness, it is revealed that sufficient angle difference between far-field Bob and Willie as well as adequate range difference between near-field Bob and Willie are necessary for ensuring covertness in fixed-position XL-array systems, while this requirement can be relaxed in movable XL-array systems thanks to flexible channel correlation control between Bobs and Willie. Next, for general system setups, we formulate an optimization problem to maximize the achievable sum-rate under covertness constraint. To solve this non-convex optimization problem, we first decompose it into two subproblems, corresponding to an inner problem for beamforming optimization given positions of subarrays and an outer problem for subarray movement optimization. Although these two subproblems are still non-convex, we obtain their high-quality solutions by using the successive convex approximation technique and devising a customized differential evolution algorithm, respectively. Last, numerical results demonstrate the effectiveness of proposed movable XL-array in balancing sum-rate and covert communication requirements.

This article presents an adaptive Super-Twisting Sliding Mode Control framework for uncertain first-order systems, with rate-bounded perturbations, where the bound is constant but unknown. Positive definite barrier functions, when used in self-tuning super-twisting controllers may introduce some conservatism in relation to initial estimations of the perturbation rate bound. Moreover, discrete time implementation of the algorithm does not necessarily guarantee the boundedness of the closed-loop trajectories when sudden changes in the perturbation occur in between two time samples. The salient features of the proposed methodology pertain to extending the use of positive semidefinite barrier functions to Super-Twisting controller adaptation and the employment of a "nested barriers" scheme that ensures boundedness of the solutions even for "unfavourable" perturbations-to-sampling time ratios. The stability of the closed-loop system is assessed via Lyapunov analysis and simulations demonstrate the efficacy of the proposed framework.

Large speech recognition models like Whisper-small achieve high accuracy but are difficult to deploy on edge devices due to their high computational demand. To this end, we present a unified, cross-library evaluation of post-training quantization (PTQ) on Whisper-small that disentangles the impact of quantization scheme, method, granularity, and bit-width. Our study is based on four libraries: PyTorch, Optimum-Quanto, HQQ, and bitsandbytes. Experiments on LibriSpeech test-clean and test-other show that dynamic int8 quantization with Quanto offers the best trade-off, reducing model size by 57% while improving on the baseline's word error rate. Static quantization performed worse, likely due to Whisper's Transformer architecture, while more aggressive formats (e.g., nf4, int3) achieved up to 71% compression at the cost of accuracy in noisy conditions. Overall, our results demonstrate that carefully chosen PTQ methods can substantially reduce model size and inference cost without retraining, enabling efficient deployment of Whisper-small on constrained hardware.

We challenge the conventional view of neural network pruning as solely a compression technique, demonstrating that one-shot magnitude pruning serves as a powerful implicit regularizer for ASR. Using Whisper-small, we combine gradient- and Fisher-based sensitivity diagnostics with targeted, component-wise pruning. This reveals architectural asymmetries: decoder FFNs are pruning-fragile, whereas decoder self-attention and the last encoder layers contain redundancy that, when removed, improves generalization. Without fine-tuning, pruning 50% of decoder self-attention reduces WER by 2.38% absolute (20.44% relative) on LibriSpeech test-other; pruning the last four encoder layers at 50% instead yields a 1.72% absolute (14.8% relative) improvement. Gains persisted on Common Voice and TED-LIUM datasets. Beyond regularization benefits, our sensitivity-aware approach enables more aggressive one-shot compression. At 40% sparsity, where established global pruning approaches catastrophically fail, our method preserves near-baseline accuracy. This positions pruning as a first-class architectural design tool: knowing where to prune is as important as how much to prune.

Frequency Modulated Continuous Wave (FMCW) radars can measure subtle chest wall oscillations to enable non-contact heartbeat sensing. However, traditional radar-based heartbeat sensing methods face performance degradation due to noise. Learning-based radar methods achieve better noise robustness but require costly labeled signals for supervised training. To overcome these limitations, we propose the first unsupervised framework for radar-based heartbeat sensing via Augmented Pseudo-Label and Noise Contrast (Radar-APLANC). We propose to use both the heartbeat range and noise range within the radar range matrix to construct the positive and negative samples, respectively, for improved noise robustness. Our Noise-Contrastive Triplet (NCT) loss only utilizes positive samples, negative samples, and pseudo-label signals generated by the traditional radar method, thereby avoiding dependence on expensive ground-truth physiological signals. We further design a pseudo-label augmentation approach featuring adaptive noise-aware label selection to improve pseudo-label signal quality. Extensive experiments on the Equipleth dataset and our collected radar dataset demonstrate that our unsupervised method achieves performance comparable to state-of-the-art supervised methods. Our code, dataset, and supplementary materials can be accessed from https://github.com/RadarHRSensing/Radar-APLANC.

Recent years have witnessed the rapid advancements of large language models (LLMs) and their expanding applications, leading to soaring demands for computational resources. The widespread adoption of test-time scaling further intensifies the tension between model capability and resource consumption. However, a rigorous metric that accurately reflects an LLM's inference efficiency across diverse tokenizers, parameter counts, and model architectures remains absent. Motivated by the correlation between compression and intelligence, we introduce information capacity, a measure of model efficiency based on text compression performance relative to computational complexity. A distinctive feature of information capacity is its incorporation of tokenizer efficiency, which affects inference costs but is often neglected in LLM evaluations. We assess the information capacity of 56 open-source models and observe a consistent information capacity among different-sized models within a series. Experiments on five heterogeneous datasets reveal strong linguistic biases in mainstream LLMs. Empirical results verify the accuracy of performance prediction across model sizes based on information capacity and show the correlation between information capacity and benchmark scores. This metric can be used to quantify improvements in inference efficiency and provide insights into better scaling performance for future LLM development.

This work introduces the Re$^{\text{2}}$MaP method, which generates expert-quality macro placements through recursively prototyping and packing tree-based relocating. We first perform multi-level macro grouping and PPA-aware cell clustering to produce a unified connection matrix that captures both wirelength and dataflow among macros and clusters. Next, we use DREAMPlace to build a mixed-size placement prototype and obtain reference positions for each macro and cluster. Based on this prototype, we introduce ABPlace, an angle-based analytical method that optimizes macro positions on an ellipse to distribute macros uniformly near chip periphery, while optimizing wirelength and dataflow. A packing tree-based relocating procedure is then designed to jointly adjust the locations of macro groups and the macros within each group, by optimizing an expertise-inspired cost function that captures various design constraints through evolutionary search. Re$^{\text{2}}$MaP repeats the above process: Only a subset of macro groups are positioned in each iteration, and the remaining macros are deferred to the next iteration to improve the prototype's accuracy. Using a well-established backend flow with sufficient timing optimizations, Re$^{\text{2}}$MaP achieves up to 22.22% (average 10.26%) improvement in worst negative slack (WNS) and up to 97.91% (average 33.97%) improvement in total negative slack (TNS) compared to the state-of-the-art academic placer Hier-RTLMP. It also ranks higher on WNS, TNS, power, design rule check (DRC) violations, and runtime than the conference version ReMaP, across seven tested cases. Our code is available at https://github.com/lamda-bbo/Re2MaP.

This paper presents a method for determining spacecraft angular rates using event-based camera sensing. This is achieved by analyzing the temporal distribution of brightness events triggered by the apparent motion of stars. The location and polarity of the events are used to infer the apparent motion field of the stars, which is, in turn, employed to estimate the observer angular velocity in the camera frame. This can be converted to the spacecraft angular rates provided an attitude reference. The method is validated through numerical simulation for a synthetic dataset of event streams generated on random spacecraft pointing and rates conditions. The accuracy of the method is assessed, demonstrating its potential to complement or replace conventional rate sensors in spacecraft systems using event camera sensing.

Music mixing involves combining individual tracks into a cohesive mixture, a task characterized by subjectivity where multiple valid solutions exist for the same input. Existing automatic mixing systems treat this task as a deterministic regression problem, thus ignoring this multiplicity of solutions. Here we introduce MEGAMI (Multitrack Embedding Generative Auto MIxing), a generative framework that models the conditional distribution of professional mixes given unprocessed tracks. MEGAMI uses a track-agnostic effects processor conditioned on per-track generated embeddings, handles arbitrary unlabeled tracks through a permutation-equivariant architecture, and enables training on both dry and wet recordings via domain adaptation. Our objective evaluation using distributional metrics shows consistent improvements over existing methods, while listening tests indicate performances approaching human-level quality across diverse musical genres.

Understanding how strategic interactions and power distributions coevolve in international relations is central to explaining conflict, cooperation, and long-term inequality. We study this problem using a power-allocation game on signed networks. Departing from models that restrict strategy updates to Pareto improvements, we propose a generalized formulation in which countries prioritize self-survival and strategically trade off between supporting allies and weakening adversaries. This relaxation allows countries to sacrifice certain allies to achieve higher overall payoffs. For the resulting static game, we establish the existence of pure-strategy Nash equilibria and characterize their properties in extreme cases, including fully antagonistic networks and the presence of a dominant power. We further introduce a power-strategy coevolutionary dynamic and prove its almost-sure convergence to equilibria corresponding to the static game. The proposed models are validated using empirical data and numerical simulations. Historical data from the Correlates of War and national capability datasets show that survival likelihood predicts countries' safety outcomes and subsequent economic growth with relatively high accuracy. Simulations further indicate that, under fixed conflict intensity, more structurally balanced signed networks yield higher average power and lower inequality at steady states.

Zhong-Ping Jiang devoted a large part of his work to the study of the stability properties of interconnected systems. In this short paper we celebrate Zhong-Ping Jiang's 60th birthday by studying a special class of families of interconnected systems: the so-called strings. We develop trajectory-based and Lyapunov-based tools that allow the verification of string stability to homogeneous bidirectional strings. The obtained results are applied to the problem of cruise controller design for a string of vehicles.

This paper presents a tutorial and survey on Probabilistic Inference-based Model Predictive Control (PI-MPC). PI-MPC reformulates finite-horizon optimal control as inference over an optimal control distribution expressed as a Boltzmann distribution weighted by a control prior, and generates actions through variational inference. In the tutorial part, we derive this formulation and explain action generation via variational inference, highlighting Model Predictive Path Integral (MPPI) control as a representative algorithm with a closed-form sampling update. In the survey part, we organize existing PI-MPC research around key design dimensions, including prior design, multi-modality, constraint handling, scalability, hardware acceleration, and theoretical analysis. This paper provides a unified conceptual perspective on PI-MPC and a practical entry point for researchers and practitioners in robotics and other control applications.

Recent implicit neural representation (INR)-based image compression methods have shown competitive performance by overfitting image-specific latent codes. However, they remain inferior to end-to-end (E2E) compression approaches due to the absence of expressive latent representations. On the other hand, E2E methods rely on transmitting latent codes and requiring complex entropy models, leading to increased decoding complexity. Inspired by the normalization strategy in E2E codecs where latents are transformed into Gaussian noise to demonstrate the removal of spatial redundancy, we explore the inverse direction: generating latents directly from Gaussian noise. In this paper, we propose a novel image compression paradigm that reconstructs image-specific latents from a multi-scale Gaussian noise tensor, deterministically generated using a shared random seed. A Gaussian Parameter Prediction (GPP) module estimates the distribution parameters, enabling one-shot latent generation via reparameterization trick. The predicted latent is then passed through a synthesis network to reconstruct the image. Our method eliminates the need to transmit latent codes while preserving latent-based benefits, achieving competitive rate-distortion performance on Kodak and CLIC dataset. To the best of our knowledge, this is the first work to explore Gaussian latent generation for learned image compression.

Robotic systems driven by conventional motors often suffer from challenges such as large mass, complex control algorithms, and the need for additional braking mechanisms, which limit their applications in lightweight and compact robotic platforms. Electrostatic film actuators offer several advantages, including thinness, flexibility, lightweight construction, and high open-loop positioning accuracy. However, the actuation stress exhibited by conventional actuators in air still needs improvement, particularly for the widely used three-phase electrode design. To enhance the output performance of actuators, this paper presents a two-layer electrostatic film actuator with an integrated brake. By alternately distributing electrodes on both the top and bottom layers, a smaller effective electrode pitch is achieved under the same fabrication constraints, resulting in an actuation stress of approximately 241~N/m$^2$, representing a 90.5% improvement over previous three-phase actuators operating in air. Furthermore, its integrated electrostatic adhesion mechanism enables load retention under braking mode. Several demonstrations, including a tug-of-war between a conventional single-layer actuator and the proposed two-layer actuator, a payload operation, a one-degree-of-freedom robotic arm, and a dual-mode gripper, were conducted to validate the actuator's advantageous capabilities in both actuation and braking modes.

Power-logistics scheduling in modern seaports typically follows a predict-then-optimize pipeline. To enhance the decision quality of predictions, decision-focused learning has been proposed, which aligns the training of forecasting models with downstream decision outcomes. However, this end-to-end design inherently restricts the value of forecasting models to a specific task structure and therefore generalizes poorly to evolving tasks induced by varying vessel arrivals. We address this gap with a decision-focused continual learning framework that adapts online to a stream of scheduling tasks. Specifically, we introduce Fisher-information-based regularization to enhance cross-task generalization by preserving parameters critical to prior tasks. A differentiable convex surrogate is also developed to stabilize gradient backpropagation. The proposed approach enables learning a decision-aligned forecasting model across a varying task stream with sustainable long-term computational and memory requirements. Experiments calibrated to Jurong Port show improved decision performance and cross-task generalization over existing methods, together with reduced computational cost and a bounded memory footprint.

A recently developed data-driven Kalman filter requires offline measurement of the process disturbance; a requirement that is often unmet for many practical applications. We propose a solution that parametrizes the Kalman filter exclusively using measured input and output data. The key idea is to use the innovations form which naturally accounts for the process disturbance and measurement noise into a single orthogonal stochastic process. Unlike process disturbances, the innovations process can be estimated directly from input-output data via a numerically efficient projection step. The performance of the method is demonstrated using a benchmark simulation.

Prevailing quantization techniques in Learned Image Compression (LIC) typically employ a static, uniform bit-width across all layers, failing to adapt to the highly diverse data distributions and sensitivity characteristics inherent in LIC models. This leads to a suboptimal trade-off between performance and efficiency. In this paper, we introduce DynaQuant, a novel framework for dynamic mixed-precision quantization that operates on two complementary levels. First, we propose content-aware quantization, where learnable scaling and offset parameters dynamically adapt to the statistical variations of latent features. This fine-grained adaptation is trained end-to-end using a novel Distance-aware Gradient Modulator (DGM), which provides a more informative learning signal than the standard Straight-Through Estimator. Second, we introduce a data-driven, dynamic bit-width selector that learns to assign an optimal bit precision to each layer, dynamically reconfiguring the network's precision profile based on the input data. Our fully dynamic approach offers substantial flexibility in balancing rate-distortion (R-D) performance and computational cost. Experiments demonstrate that DynaQuant achieves rd performance comparable to full-precision models while significantly reducing computational and storage requirements, thereby enabling the practical deployment of advanced LIC on diverse hardware platforms.

Safety assurance is a fundamental requirement for deploying learning-enabled autonomous systems. Hamilton-Jacobi (HJ) reachability analysis is a fundamental method for formally verifying safety and generating safe controllers. However, computing the HJ value function that characterizes the backward reachable set (BRS) of a set of user-defined failure states is computationally expensive, especially for high-dimensional systems, motivating the use of reinforcement learning approaches to approximate the value function. Unfortunately, a learned value function and its corresponding safe policy are not guaranteed to be correct. The learned value function evaluated at a given state may not be equal to the actual safety return achieved by following the learned safe policy. To address this challenge, we introduce a conformal prediction-based (CP) framework that bounds such uncertainty. We leverage CP to provide probabilistic safety guarantees when using learned HJ value functions and policies to prevent control systems from reaching failure states. Specifically, we use CP to calibrate the switching between the unsafe nominal controller and the learned HJ-based safe policy and to derive safety guarantees under this switched policy. We also investigate using an ensemble of independently trained HJ value functions as a safety filter and compare this ensemble approach to using individual value functions alone.