Papers

This paper introduces a novel methodology of adapting the representation of videos based on the dynamics of their scene content variation. In particular, we demonstrate how the clustering of dynamic mode decomposition eigenvalues can be leveraged to learn an adaptive video representation for separating structurally distinct morphologies of a video. We extend the morphological component analysis (MCA) algorithm, which uses multiple predefined incoherent dictionaries and a sparsity prior to separate distinct sources in signals, by introducing our novel eigenspace clustering technique to obtain data-driven MCA dictionaries, which we call dynamic morphological component analysis (DMCA). After deriving our novel algorithm, we offer a motivational example of DMCA applied to a still image, then demonstrate DMCA's effectiveness in denoising applications on videos from the Adobe 240fps dataset. Afterwards, we provide an example of DMCA enhancing the signal-to-noise ratio of a faint target summed with a sea state, and conclude the paper by applying DMCA to separate a bicycle from wind clutter in inverse synthetic aperture radar images.

Distributed control of connected and automated vehicles has attracted considerable interest for its potential to improve traffic efficiency and safety. However, such control schemes require sharing privacy-sensitive vehicle data, which introduces risks of information leakage and potential malicious activities. This paper investigates the stability and privacy-preserving properties of distributed platoon control under two types of quantizers: deterministic and probabilistic. For deterministic quantization, we show that the resulting control strategy ensures the system errors remain uniformly ultimately bounded. Moreover, in the absence of auxiliary information, an eavesdropper cannot uniquely infer sensitive vehicle states. In contrast, the use of probabilistic quantization enables asymptotic convergence of the vehicle platoon in expectation with bounded variance. Importantly, probabilistic quantizers can satisfy differential privacy guarantees, thereby preserving privacy even when the eavesdropper possesses arbitrary auxiliary information. We further analyze the trade-off between control performance and privacy by formulating an optimization problem that characterizes the impact of the quantization step on both metrics. Numerical simulations are provided to illustrate the performance differences between the two quantization strategies.

Over the past two decades, speech emotion recognition (SER) has received growing attention. To train SER systems, researchers collect emotional speech databases annotated by crowdsourced or in-house raters who select emotions from predefined categories. However, disagreements among raters are common. Conventional methods treat these disagreements as noise, aggregating labels into a single consensus target. While this simplifies SER as a single-label task, it ignores the inherent subjectivity of human emotion perception. This dissertation challenges such assumptions and asks: (1) Should minority emotional ratings be discarded? (2) Should SER systems learn from only a few individuals' perceptions? (3) Should SER systems predict only one emotion per sample? Psychological studies show that emotion perception is subjective and ambiguous, with overlapping emotional boundaries. We propose new modeling and evaluation perspectives: (1) Retain all emotional ratings and represent them with soft-label distributions. Models trained on individual annotator ratings and jointly optimized with standard SER systems improve performance on consensus-labeled tests. (2) Redefine SER evaluation by including all emotional data and allowing co-occurring emotions (e.g., sad and angry). We propose an “all-inclusive rule” that aggregates all ratings to maximize diversity in label representation. Experiments on four English emotion databases show superior performance over majority and plurality labeling. (3) Construct a penalization matrix to discourage unlikely emotion combinations during training. Integrating it into loss functions further improves performance. Overall, embracing minority ratings, multiple annotators, and multi-emotion predictions yields more robust and human-aligned SER systems.

Mel-frequency cepstral coefficients (MFCCs) are an important feature in speech processing. A deeper understanding of their properties can contribute to the work that is being done with both classical and deep learning models. This study challenges the long-held assumption that MFCCs lack relevant temporal information by investigating their relationship with speech prosody. Using a null hypothesis significance testing framework, a systematic assessment is made about the statistical independence between MFCCs and the three prosodic features: energy, fundamental frequency (F0), and voicing. The results demonstrate that it is statistically implausible that the MFCCs are independent of any of these three prosodic features. This finding suggests that MFCCs inherently carry valuable prosodic information, which can inform the design of future models in speech analysis and recognition.

Soft landing on small celestial bodies (SCBs) poses unique challenges, as gravitational models poorly characterize the higher-order gravitational effects of SCBs. Existing control approaches lack guarantees for safety under gravitational uncertainty. This paper proposes a three-stage control architecture that combines disturbance estimation, trajectory tracking, and safety enforcement. An extended high-gain observer estimates gravitational disturbances online, a feedback-linearizing controller tracks a reference trajectory, and a minimum-intervention quadratic program enforces state and input constraints while remaining close to the nominal control. The proposed approach enables aggressive yet safe maneuvers despite gravitational uncertainty. Numerical simulations demonstrate the effectiveness of the controller in achieving soft-landing on irregularly shaped SCBs, highlighting its potential for autonomous SCB missions.

We propose the Segmented Full-Song Model (SFS) for symbolic full-song generation. The model accepts a user-provided song structure and an optional short seed segment that anchors the main idea around which the song is developed. By factorizing a song into segments and generating each one through selective attention to related segments, the model achieves higher quality and efficiency compared to prior work. To demonstrate its suitability for human-AI interaction, we further wrap SFS into a web application that enables users to iteratively co-create music on a piano roll with customizable structures and flexible ordering.

This paper presents a framework for time-causal wavelet analysis. It targets real-time processing of temporal signals, where data from the future are not available. The study builds upon temporal scale-space theory, originating from a complete classification of temporal smoothing kernels that guarantee non-creation of new structures from finer to coarser temporal scale levels. We construct temporal wavelets from the temporal derivatives of a special time-causal smoothing kernel, referred to as the time-causal limit kernel, as arising from the classification of variation-diminishing smoothing transformations with the complementary requirement of temporal scale covariance, to guarantee self-similar handling of structures in the input signal at different temporal scales. This enables decomposition of the signal into different components at different scales, while adhering to temporal causality. The paper establishes theoretical foundations for these time-causal wavelet representations, and maps structural relationships to the non-causal Ricker or Mexican hat wavelets. We also describe how efficient discrete approximations of the presented theory can be performed in terms of first-order recursive filters coupled in cascade, which enables numerically well-conditioned real-time processing with low resource usage. We characterize and quantify how the continuous scaling properties transfer to the discrete implementation, demonstrating how the proposed time-causal wavelet representation can reflect the duration of locally dominant temporal structures in the input signal.

In this work, we present FoleyGRAM, a novel approach to video-to-audio generation that emphasizes semantic conditioning through the use of aligned multimodal encoders. Building on prior advancements in video-to-audio generation, FoleyGRAM leverages the Gramian Representation Alignment Measure (GRAM) to align embeddings across video, text, and audio modalities, enabling precise semantic control over the audio generation process. The core of FoleyGRAM is a diffusion-based audio synthesis model conditioned on GRAM-aligned embeddings and waveform envelopes, ensuring both semantic richness and temporal alignment with the corresponding input video. We evaluate FoleyGRAM on the Greatest Hits dataset, a standard benchmark for video-to-audio models. Our experiments demonstrate that aligning multimodal encoders using GRAM enhances the system's ability to semantically align generated audio with video content, advancing the state of the art in video-to-audio synthesis.

Although audio generation has been widely studied over recent years, video-aligned audio generation still remains a relatively unexplored frontier. To address this gap, we introduce StereoSync, a novel and efficient model designed to generate audio that is both temporally synchronized with a reference video and spatially aligned with its visual context. Moreover, StereoSync also achieves efficiency by leveraging pretrained foundation models, reducing the need for extensive training while maintaining high-quality synthesis. Unlike existing methods that primarily focus on temporal synchronization, StereoSync introduces a significant advancement by incorporating spatial awareness into video-aligned audio generation. Indeed, given an input video, our approach extracts spatial cues from depth maps and bounding boxes, using them as cross-attention conditioning in a diffusion-based audio generation model. Such an approach allows StereoSync to go beyond simple synchronization, producing stereo audio that dynamically adapts to the spatial structure and movement of a video scene. We evaluate StereoSync on Walking The Maps, a curated dataset comprising videos from video games that feature animated characters walking through diverse environments. Experimental results demonstrate the ability of StereoSync to achieve both temporal and spatial alignment, advancing the state of the art in video-to-audio generation and resulting in a significantly more immersive and realistic audio experience.

Biomedical signals provide insights into various conditions affecting the human body. Beyond diagnostic capabilities, these signals offer a deeper understanding of how specific organs respond to an individual's emotions and feelings. For instance, ECG data can reveal changes in heart rate variability linked to emotional arousal, stress levels, and autonomic nervous system activity. This data offers a window into the physiological basis of our emotional states. Recent advancements in the field diverge from conventional approaches by leveraging the power of advanced transformer architectures, which surpass traditional machine learning and deep learning methods. We begin by assessing the effectiveness of the Vision Transformer (ViT), a forefront model in image classification, for identifying emotions in imaged ECGs. Following this, we present and evaluate an improved version of ViT, integrating both CNN and SE blocks, aiming to bolster performance on imaged ECGs associated with emotion detection. Our method unfolds in two critical phases: first, we apply advanced preprocessing techniques for signal purification and converting signals into interpretable images using continuous wavelet transform and power spectral density analysis; second, we unveil a performance-boosted vision transformer architecture, cleverly enhanced with convolutional neural network components, to adeptly tackle the challenges of emotion recognition. Our methodology's robustness and innovation were thoroughly tested using ECG data from the YAAD and DREAMER datasets, leading to remarkable outcomes. For the YAAD dataset, our approach outperformed existing state-of-the-art methods in classifying seven unique emotional states, as well as in valence and arousal classification. Similarly, in the DREAMER dataset, our method excelled in distinguishing between valence, arousal and dominance, surpassing current leading techniques.

Wearable robots offer a promising solution for quantitatively monitoring gait and providing systematic, adaptive assistance to promote patient independence and improve gait. However, due to significant interpersonal and intrapersonal variability in walking patterns, it is important to design robot controllers that can adapt to the unique characteristics of each individual. This paper investigates the potential of human-in-the-loop optimisation (HILO) to deliver personalised assistance in gait training. The Covariance Matrix Adaptation Evolution Strategy (CMA-ES) was employed to continuously optimise an assist-as-needed controller of a lower-limb exoskeleton. Six healthy individuals participated over a two-day experiment. Our results suggest that while the CMA-ES appears to converge to a unique set of stiffnesses for each individual, no measurable impact on the subjects' performance was observed during the validation trials. These findings highlight the impact of human-robot co-adaptation and human behaviour variability, whose effect may be greater than potential benefits of personalising rule-based assistive controllers. Our work contributes to understanding the limitations of current personalisation approaches in exoskeleton-assisted gait rehabilitation and identifies key challenges for effective implementation of human-in-the-loop optimisation in this domain.

We address monaural multi-speaker-image separation in reverberant conditions, aiming at separating mixed speakers but preserving the reverberation of each speaker. A straightforward approach for this task is to directly train end-to-end DNN systems to predict the reverberant speech of each speaker based on the input mixture. Although effective, this approach does not explicitly exploit the physical constraint that reverberant speech can be reproduced by convolving the direct-path signal with a linear filter. To address this, we propose CxNet, a two-DNN system with a neural forward filtering module in between. The first DNN is trained to jointly predict the direct-path signal and reverberant speech. Based on the direct-path estimate, the neural forward filtering module estimates the linear filter, and the estimated filter is then convolved with the direct-path estimate to obtain another estimate of reverberant speech, which is utilized as a discriminative feature to help the second DNN better estimate the reverberant speech. By explicitly modeling the linear filter, CxNet could leverage the physical constraint between the direct-path signal and reverberant speech to capture crucial information about reverberation tails. Evaluation results on the SMS-WSJ dataset show the effectiveness of the proposed algorithms.

Whereas chord transcription has received considerable attention during the past couple of decades, far less work has been devoted to transcribing and encoding the rhythmic patterns that occur in a song. The topic is especially relevant for instruments such as the rhythm guitar, which is typically played by strumming rhythmic patterns that repeat and vary over time. However, in many cases one cannot objectively define a single "right" rhythmic pattern for a given song section. To create a dataset with well-defined ground-truth labels, we asked expert musicians to transcribe the rhythmic patterns in 410 popular songs and record cover versions where the guitar tracks followed those transcriptions. To transcribe the strums and their corresponding rhythmic patterns, we propose a three-step framework. Firstly, we perform approximate stem separation to extract the guitar part from the polyphonic mixture. Secondly, we detect individual strums within the separated guitar audio, using a pre-trained foundation model (MERT) as a backbone. Finally, we carry out a pattern-decoding process in which the transcribed sequence of guitar strums is represented by patterns drawn from an expert-curated vocabulary. We show that it is possible to transcribe the rhythmic patterns of the guitar track in polyphonic music with quite high accuracy, producing a representation that is human-readable and includes automatically detected bar lines and time signature markers. We perform ablation studies and error analysis and propose a set of evaluation metrics to assess the accuracy and readability of the predicted rhythmic pattern sequence.

Ultrafast ultrasound imaging enables visualization of rapid physiological dynamics by acquiring data at exceptionally high frame rates. However, this speed often comes at the cost of spatial resolution and image quality due to unfocused wave transmissions and associated artifacts. In this work, we propose a novel modulated Implicit Neural Representation (INR) framework that leverages a coordinate-based neural network conditioned on latent embeddings extracted from time-delayed I/Q channel data for high-quality ultrasound image reconstruction. Our method integrates complex Gabor wavelet activation and a conditioner network to capture the oscillatory and phase-sensitive nature of I/Q ultrasound signals. We evaluate the framework on an in vivo intracardiac echocardiography (ICE) dataset and demonstrate that it outperforms the compared state-of-the-art methods. We believe these findings not only highlight the advantages of INR-based modeling for ultrasound image reconstruction, but also point to broader opportunities for applying INR frameworks across other medical imaging modalities.

Given the speech generation framework that represents the speaker attribute with an embedding vector, asynchronous voice anonymization can be achieved by modifying the speaker embedding derived from the original speech. However, the inconsistency between machine and human perceptions of the speaker attribute within the speaker embedding remains unexplored, limiting its performance in asynchronous voice anonymization. To this end, this study investigates this inconsistency via modifications to speaker embedding in the speech generation process. Experiments conducted on the FACodec and Diff-HierVC speech generation models discover a subspace whose removal alters machine perception while preserving its human perception of the speaker attribute in the generated speech. With these findings, an asynchronous voice anonymization is developed, achieving 100% human perception preservation rate while obscuring the machine perception. Audio samples can be found in https://voiceprivacy.github.io/speaker-embedding-eigen-decomposition/.

Federated split learning (FedSL) has emerged as a promising paradigm for enabling collaborative intelligence in industrial Internet of Things (IoT) systems, particularly in smart factories where data privacy, communication efficiency, and device heterogeneity are critical concerns. In this article, we present a comprehensive study of FedSL frameworks tailored for resource-constrained robots in industrial scenarios. We compare synchronous, asynchronous, hierarchical, and heterogeneous FedSL frameworks in terms of workflow, scalability, adaptability, and limitations under dynamic industrial conditions. Furthermore, we systematically categorize token fusion strategies into three paradigms: input-level (pre-fusion), intermediate-level (intra-fusion), and output-level (post-fusion), and summarize their respective strengths in industrial applications. We also provide adaptive optimization techniques to enhance the efficiency and feasibility of FedSL implementation, including model compression, split layer selection, computing frequency allocation, and wireless resource management. Simulation results validate the performance of these frameworks under industrial detection scenarios. Finally, we outline open issues and research directions of FedSL in future smart manufacturing systems.

Cross-entropy method model predictive control (CEM–MPC) is a powerful gradient-free technique for nonlinear optimal control, but its performance is often limited by the reliance on random sampling. This conventional approach can lead to inefficient exploration of the solution space and non-smooth control inputs, requiring a large number of samples to achieve satisfactory results. To address these limitations, we propose deterministic sampling CEM (dsCEM), a novel framework that replaces the random sampling step with deterministic samples derived from localized cumulative distributions (LCDs). Our approach introduces modular schemes to generate and adapt these sample sets, incorporating temporal correlations to ensure smooth control trajectories. This method can be used as a drop-in replacement for the sampling step in existing CEM-based controllers. Experimental evaluations on two nonlinear control tasks demonstrate that dsCEM consistently outperforms state-of-the-art iCEM in terms of cumulative cost and control input smoothness, particularly in the critical low-sample regime.

Ultrafast intracardiac echocardiography (ICE) uses unfocused transmissions to capture cardiac motion at frame rates exceeding 1 kHz. While this enables real-time visualization of rapid dynamics, image quality is often degraded by diffraction artifacts, requiring many transmits to achieve satisfying resolution and contrast. To address this limitation, we propose an implicit neural representation (INR) framework to encode complex-valued receive apodization weights in a continuous manner, enabling high-quality ICE reconstructions from only three diverging wave (DW) transmits. Our method employs a multi-layer perceptron that maps pixel coordinates and transmit steering angles to complex-valued apodization weights for each receive channel. Experiments on a large in vivo porcine ICE imaging dataset show that the learned apodization suppresses clutter and enhances contrast, yielding reconstructions closely matching 26-angle compounded DW ground truths. Our study suggests that INRs could offer a powerful framework for ultrasound image enhancement.

The rapid development of Generative Artificial Intelligence (GenAI) has catalyzed a transformative technological revolution across all walks of life. As the backbone of wideband communication, optical networks are expecting high-level autonomous operation and zero-touch management to accommodate their expanding network scales and escalating transmission bandwidth. The integration of GenAI is deemed as the pivotal solution for realizing zero-touch optical networks. However, the lifecycle management of optical networks involves a multitude of tasks and necessitates seamless collaboration across multiple layers, which poses significant challenges to the existing single-agent GenAI systems. In this paper, we propose a GenAI-driven hierarchical multi-agent framework designed to streamline multi-task autonomous execution for zero-touch optical networks. We present the architecture, implementation, and applications of this framework. A field-deployed mesh network is utilized to demonstrate three typical scenarios throughout the lifecycle of optical network: quality of transmission estimation in the planning stage, dynamic channel adding/dropping in the operation stage, and system capacity increase in the upgrade stage. The case studies, illustrate the capabilities of multi-agent framework in multi-task allocation, coordination, execution, evaluation, and summarization. This work provides a promising approach for the future development of intelligent, efficient, and collaborative network management solutions, paving the way for more specialized and adaptive zero-touch optical networks.

Current speech production systems predominantly rely on large transformer models that operate as black boxes, providing little interpretability or grounding in the physical mechanisms of human speech. We address this limitation by proposing a new framework: speech generation through explicit articulatory control. This reframes speech as a motor control task similar to robotic manipulation. Our approach uses reinforcement learning to train a policy that directly controls the movements of vocal tract articulators, such as the tongue, lips, and jaw, to produce syllable-level speech. Specifically, we employ the Proximal Policy Optimization algorithm to learn optimal articulatory movements based on acoustic feedback provided by our audio perceiver, Sylber. The resulting articulatory trajectories are decoded into audio using SPARC, a pre-trained articulatory-to-speech decoder. We train this framework on six target syllables, and it demonstrates successful convergence, with similarity scores between the policy-generated audio and the target syllables exceeding 0.85. Accurate human transcription of the audio for syllables such as "please", "loot", and "cat" demonstrates the intelligibility of this framework.

Statistical significance testing of neural coherence is essential for distinguishing genuine cross-signal coupling from spurious correlations. A widely accepted approach uses surrogate-based inference, where null distributions are generated via time-shift or phase-randomization procedures. While effective, these methods are computationally expensive and yield discrete p-values that can be unstable near decision thresholds, limiting scalability to large EEG/iEEG datasets. We introduce and validate a parametric alternative based on a generalized linear model (GLM) applied to complex-valued time–frequency coefficients (e.g., from DBT or STFT), using a likelihood-ratio test. Using real respiration belt traces as a driver and simulated neural signals contaminated with broadband Gaussian noise, we perform dense sweeps of ground-truth coherence and compare GLM-based inference against time-shift/phase-randomized surrogate testing under matched conditions. GLM achieved comparable or superior sensitivity while producing continuous, stable p-values and a substantial computational advantage. At 80% detection power, GLM detects at C=0.25, whereas surrogate testing requires C=0.49, corresponding to an approximately 6–7 dB SNR improvement. Runtime benchmarking showed GLM to be nearly 200x faster than surrogate approaches. These results establish GLM-based inference on complex time–frequency coefficients as a robust, scalable alternative to surrogate testing, enabling efficient analysis of large EEG/iEEG datasets across channels, frequencies, and participants.

Purpose: To develop and validate No-New SAM2 (nnsam2) for few-shot segmentation of lumbar paraspinal muscles using only a single annotated slice per dataset, and to assess its statistical comparability with expert measurements across multi-sequence MRI and multi-protocol CT. Methods: We retrospectively analyzed 1,219 scans (19,439 slices) from 762 participants across six datasets. Six slices (one per dataset) served as labeled examples, while the remaining 19,433 slices were used for testing. In this minimal-supervision setting, nnsam2 used single-slice SAM2 prompts to generate pseudo-labels, which were pooled across datasets and refined through three sequential, independent nnU-Net models. Segmentation performance was evaluated using the Dice similarity coefficient (DSC), and automated measurements-including muscle volume, fat ratio, and CT attenuation-were assessed with two one-sided tests (TOST) and intraclass correlation coefficients (ICC). Results: nnsam2 outperformed vanilla SAM2, its medical variants, TotalSegmentator, and the leading few-shot method, achieving DSCs of 0.94-0.96 on MR images and 0.92-0.93 on CT. Automated and expert measurements were statistically equivalent for muscle volume (MRI/CT), CT attenuation, and Dixon fat ratio (TOST, P < 0.05), with consistently high ICCs (0.86-1.00). Conclusion: We developed nnsam2, a state-of-the-art few-shot framework for multi-modality LPM segmentation, producing muscle volume (MRI/CT), attenuation (CT), and fat ratio (Dixon MRI) measurements that were statistically comparable to expert references. Validated across multimodal, multicenter, and multinational cohorts, and released with open code and data, nnsam2 demonstrated high annotation efficiency, robust generalizability, and reproducibility.

Artificial Potential Field (APF) methods are widely used for reactive flocking control, but they often suffer from challenges such as deadlocks and local minima, especially in the presence of obstacles. Existing solutions to address these issues are typically passive, leading to slow and inefficient collective navigation. As a result, many APF approaches have only been validated in obstacle-free environments or simplified, pseudo 3D simulations. This paper presents GO-Flock, a hybrid flocking framework that integrates planning with reactive APF-based control. GO-Flock consists of an upstream Perception Module, which processes depth maps to extract waypoints and virtual agents for obstacle avoidance, and a downstream Collective Navigation Module, which applies a novel APF strategy to achieve effective flocking behavior in cluttered environments. We evaluate GO-Flock against passive APF-based approaches to demonstrate their respective merits, such as their flocking behavior and the ability to overcome local minima. Finally, we validate GO-Flock through obstacle-filled environment and also hardware-in-the-loop experiments where we successfully flocked a team of nine drones, six physical and three virtual, in a forest environment.

Acoustic scene perception involves describing the type of sounds, their timing, their direction and distance, as well as their loudness and reverberation. While audio language models excel in sound recognition, single-channel input fundamentally limits spatial understanding. This work presents Sci-Phi, a spatial audio large language model with dual spatial and spectral encoders that estimates a complete parameter set for all sound sources and the surrounding environment. Learning from over 4,000 hours of synthetic first-order Ambisonics recordings including metadata, Sci-Phi enumerates and describes up to four directional sound sources in one pass, alongside non-directional background sounds and room characteristics. We evaluate the model with a permutation-invariant protocol and 15 metrics covering content, location, timing, loudness, and reverberation, and analyze its robustness across source counts, signal-to-noise ratios, reverberation levels, and challenging mixtures of acoustically, spatially, or temporally similar sources. Notably, Sci-Phi generalizes to real room impulse responses with only minor performance degradation. Overall, this work establishes the first audio LLM capable of full spatial-scene description, with strong potential for real-world deployment. Demo: https://sci-phi-audio.github.io/demo

Large Audio Language Models (LALMs) exhibit strong capabilities in general audio understanding but remain static after deployment, limiting their adaptability to real-world data. Since supervised fine-tuning is costly, we propose AQA-TTRL, a novel framework for audio understanding that enables on-the-fly evolution via test-time reinforcement learning using only unlabeled test data. It generates pseudo-labels via majority voting and optimizes the model through reinforcement learning. To address the noise in self-generated labels, we introduce confidence weighting to adjust training signals. Furthermore, multiple-attempt sampling mitigates advantage collapse and stabilizes training. Across MMAU, MMAR, and MMSU, AQA-TTRL achieves significant average improvements of 4.42% for Qwen2.5-Omni 7B and 11.04% for the 3B model. Notably, the adapted 3B model outperforms direct inference of the unadapted 7B model, highlighting the effectiveness of test-time adaptation in audio understanding.

Control problems frequently arise in scientific and industrial applications, where the objective is to steer a dynamical system from an initial state to a desired target state. Recent advances in deep learning and automatic differentiation have made applying these methods to control problems increasingly practical. In this paper, we examine the use of neural networks and modern machine-learning libraries to parameterize control inputs across discrete-time and continuous-time systems, as well as deterministic and stochastic dynamics. We highlight applications in multiple domains, including biology, engineering, physics, and medicine. For continuous-time dynamical systems, neural ordinary differential equations (neural ODEs) offer a useful approach to parameterizing control inputs. For discrete-time systems, we show how custom control-input parameterizations can be implemented and optimized using automatic-differentiation methods. Overall, the methods presented provide practical solutions for control tasks that are computationally demanding or analytically intractable, making them valuable for complex real-world applications.

Pruning large pre-trained transformers in a data-scarce scenario is challenging, as it often requires massive retraining data to recover performance. For instance, Distill-Whisper prunes Whisper by 40 and retrains on 21,000 hours of speech, far beyond what is available for most languages. Can Whisper be made lighter and faster for edge devices in data-scarce settings? Focusing on Bambara with only 32h of speech-to-text data, we propose a new pruning recipe. Instead of vocabulary pruning, which is unsuitable due to frequent code-switching by Bambara speakers, we compress the embeddings with low-rank decomposition and feature distillation. Rather than removing layers, we merge them to limit performance loss. The final model preserves 90 of the original performance while being 48 smaller and 2.15x faster on a MacBook Air M1.

Timely resource allocation in edge-assisted vehicular networks is essential for compute-intensive services such as autonomous driving and navigation. However, vehicle mobility leads to spatio-temporal unpredictability of resource demands, while real-time double auctions incur significant latency. To address these challenges, we propose a look-ahead contract-based auction framework that shifts decision-making from runtime to planning time. Our approach establishes N-step service contracts between edge servers (ESs) using demand forecasts and modified double auctions. The system operates in two stages: first, an LSTM-based prediction module forecasts multi-slot resource needs and determines ES roles (buyer or seller), after which a pre-double auction generates contracts specifying resource quantities, prices, and penalties. Second, these contracts are enforced in real time without rerunning auctions. The framework incorporates energy costs, transmission overhead, and contract breach risks into utility models, ensuring truthful, rational, and energy-efficient trading. Experiments on real-world (UTD19) and synthetic traces demonstrate that our method improves time efficiency, energy use, and social welfare compared with existing baselines.

Incremental brain tumor segmentation is critical for models that must adapt to evolving clinical datasets without retraining on all prior data. However, catastrophic forgetting, where models lose previously acquired knowledge, remains a major obstacle. Recent incremental learning frameworks with knowledge distillation partially mitigate forgetting but rely heavily on generative replay or auxiliary storage. Meanwhile, diffusion models have proven effective for refining tumor segmentations, but have not been explored in incremental learning contexts. We propose Synthetic Error Replay Diffusion (SER-Diff), the first framework that unifies diffusion-based refinement with incremental learning. SER-Diff leverages a frozen teacher diffusion model to generate synthetic error maps from past tasks, which are replayed during training on new tasks. A dual-loss formulation combining Dice loss for new data and knowledge distillation loss for replayed errors ensures both adaptability and retention. Experiments on BraTS2020, BraTS2021, and BraTS2023 demonstrate that SER-Diff consistently outperforms prior methods. It achieves the highest Dice scores of 95.8%, 94.9%, and 94.6%, along with the lowest HD95 values of 4.4 mm, 4.7 mm, and 4.9 mm, respectively. These results indicate that SER-Diff not only mitigates catastrophic forgetting but also delivers more accurate and anatomically coherent segmentations across evolving datasets.

Focal Cortical Dysplasia (FCD) is a primary cause of drug-resistant epilepsy and is difficult to detect in brain {magnetic resonance imaging} (MRI) due to the subtle and small-scale nature of its lesions. Accurate segmentation of FCD regions in 3D multimodal brain MRI images is essential for effective surgical planning and treatment. However, this task remains highly challenging due to the limited availability of annotated FCD datasets, the extremely small size and weak contrast of FCD lesions, the complexity of handling 3D multimodal inputs, and the need for output smoothness and anatomical consistency, which is often not addressed by standard voxel-wise loss functions. This paper presents a new framework for segmenting FCD regions in 3D brain MRI images. We adopt state-of-the-art transformer-enhanced encoder-decoder architecture and introduce a novel loss function combining Dice loss with an anisotropic {Total Variation} (TV) term. This integration encourages spatial smoothness and reduces false positive clusters without relying on post-processing. The framework is evaluated on a public FCD dataset with 85 epilepsy patients and demonstrates superior segmentation accuracy and consistency compared to standard loss formulations. The model with the proposed TV loss shows an 11.9% improvement on the Dice coefficient and 13.3% higher precision over the baseline model. Moreover, the number of false positive clusters is reduced by 61.6%