Scaling model parameters has become the de facto strategy for improving NLP systems, but it comes with substantial computational costs. Test-Time Scaling (TTS) offers an alternative by allocating more computation at inference: generating multiple candidates and selecting the best. While effective in tasks such as mathematical reasoning, TTS has not been systematically explored for machine translation (MT). In this paper, we present the first systematic study of TTS for MT, investigating a simple but practical best-of-N framework on WMT24 benchmarks. Our experiments cover six high-resource and one low-resource language pairs, five model sizes (3B-72B), and various TTS compute budget (N up to 1024). Our results show that a) For high-resource languages, TTS generally improves translation quality according to multiple neural MT evaluation metrics, and our human evaluation confirms these gains; b) Augmenting smaller models with large $N$ can match or surpass larger models at $N{=}1$ with more compute cost; c) Under fixed compute budgets, larger models are typically more efficient, and TTS can degrade quality due to metric blind spots in low-resource cases.

Recent advances in multimodal large language models (LLMs) have led to significant progress in understanding, generation, and retrieval tasks. However, current solutions often treat these tasks in isolation or require training LLMs from scratch, resulting in high computational costs and limited generalization across modalities. In this work, we present OmniBridge, a unified and modular multimodal framework that supports vision-language understanding, generation, and retrieval within a unified architecture. OmniBridge adopts a language-centric design that reuses pretrained LLMs and introduces a lightweight bidirectional latent alignment module. To address the challenge of task interference, we propose a two-stage decoupled training strategy: supervised fine-tuning and latent space alignment for aligning LLM behavior with multimodal reasoning, and semantic-guided diffusion training to align cross-modal latent spaces via learnable query embeddings. Extensive experiments across a wide range of benchmarks demonstrate that OmniBridge achieves competitive or state-of-the-art performance in all three tasks. Moreover, our results highlight the effectiveness of latent space alignment for unifying multimodal modeling under a shared representation space. Code and models are released at https://github.com/xiao-xt/OmniBridge.

Non-Markovian Reinforcement Learning (RL) tasks present significant challenges, as agents must reason over entire trajectories of state-action pairs to make optimal decisions. A common strategy to address this is through symbolic formalisms, such as Linear Temporal Logic (LTL) or automata, which provide a structured way to express temporally extended objectives. However, these approaches often rely on restrictive assumptions -- such as the availability of a predefined Symbol Grounding (SG) function mapping raw observations to high-level symbolic representations, or prior knowledge of the temporal task. In this work, we propose a fully learnable version of Neural Reward Machines (NRM), which can learn both the SG function and the automaton end-to-end, removing any reliance on prior knowledge. Our approach is therefore as easily applicable as classic deep RL (DRL) approaches, while being far more explainable, because of the finite and compact nature of automata. Furthermore, we show that by integrating Fully Learnable Reward Machines (FLNRM) with DRL, our method outperforms previous approaches based on Recurrent Neural Networks (RNNs).

Low-rank architectures have become increasingly important for efficient large language model (LLM) pre-training, providing substantial reductions in both parameter complexity and memory/computational demands. Despite these advantages, current low-rank methods face three critical shortcomings: (1) compromised model performance, (2) considerable computational overhead, and (3) limited activation memory savings. To address these limitations, we propose Cross-layer Low-Rank residual Network (CR-Net), an innovative parameter-efficient framework inspired by our discovery that inter-layer activation residuals possess low-rank properties. CR-Net implements this insight through a dual-path architecture that efficiently reconstructs layer activations by combining previous-layer outputs with their low-rank differences, thereby maintaining high-rank information with minimal parameters. We further develop a specialized activation recomputation strategy tailored for CR-Net that dramatically reduces memory requirements. Extensive pre-training experiments across model scales from 60M to 7B parameters demonstrate that CR-Net consistently outperforms state-of-the-art low-rank frameworks while requiring fewer computational resources and less memory.

Predictive process monitoring is a sub-domain of process mining which aims to forecast the future of ongoing process executions. One common prediction target is the remaining time, meaning the time that will elapse until a process execution is completed. In this paper, we compare four different remaining time prediction approaches in a real-life outbound warehouse process of a logistics company in the aviation business. For this process, the company provided us with a novel and original event log with 169,523 traces, which we can make publicly available. Unsurprisingly, we find that deep learning models achieve the highest accuracy, but shallow methods like conventional boosting techniques achieve competitive accuracy and require significantly fewer computational resources.

However, such energy advantage is often unrealized because inference requires evaluating a temporal decay function and subsequent multiplication with the synaptic weights. We fabricated a custom indium oxide optoelectronic synapse, showing how its natural physical decay directly implements the required temporal function. By treating the device's analog output as the fused product of the synaptic weight and temporal decay, optoelectronic synaptic TTFS (named Otters) eliminates these expensive digital operations. To use the Otters paradigm in complex architectures like the transformer, which are challenging to train directly due to the sparsity issue, we introduce a novel quantized neural network-to-SNN conversion algorithm. This complete hardware-software co-design enables our model to achieve state-of-the-art accuracy across seven GLUE benchmark datasets and demonstrates a 1.77 improvement in energy efficiency over previous leading SNNs, based on a comprehensive analysis of compute, data movement, and memory access costs using energy measurements from a commercial 22nm process. Our work thus establishes a new paradigm for energy-efficient SNNs, translating fundamental device physics directly into powerful computational primitives. Large language models (LLMs) have demonstrated remarkable capabilities, yet their immense computational and energy costs hinder their deployment in resource-constrained environments such as edge devices (Lin et al., 2023; Jegham et al., 2025). This critical challenge has spurred research on more efficient, brain-inspired architectures, with spiking neural networks (SNNs) emerging as a promising candidate (Tang et al., 2025; Xing et al.).

Due to their robustness and ability to effectively differentiate between target and background, discriminative approaches have become the leading techniques in object identification and tracking. The computational capabilities of GPUs allowed the transition to DL-based detection, leading to significant breakthroughs in object detection, improving the overall performances [19] [20]. Object detection algorithms commonly used in medicine leverage CNNs for their ability to automatically extract features from extensive training datasets, enabling accurate predictions. This has led to the development of various CNN-based methods in recent years for identifying and tracking laparoscopic instruments and organs, especially during Minimally Invasive Surgery (MIS) [21]. As reported in literature [22], Deep Convolutional Neural Network (DCNN)-based detectors can be categorized as either "two-stage" or "one-stage". Two-stage detectors separate object localization (generating region proposals) from object classification, whilst one-stage detectors perform both detection and classification simultaneously using DCNNs.

LiDAR-based localization and SLAM often rely on iterative matching algorithms, particularly the Iterative Closest Point (ICP) algorithm, to align sensor data with pre-existing maps or previous scans. However, ICP is prone to errors in featureless environments and dynamic scenes, leading to inaccurate pose estimation. Accurately predicting the uncertainty associated with ICP is crucial for robust state estimation but remains challenging, as existing approaches often rely on handcrafted models or simplified assumptions. Moreover, a few deep learning-based methods for localizability estimation either depend on a pre-built map, which may not always be available, or provide a binary classification of localizable versus non-localizable, which fails to properly model uncertainty. In this work, we propose a data-driven framework that leverages deep learning to estimate the registration error covariance of ICP before matching, even in the absence of a reference map. By associating each LiDAR scan with a reliable 6-DoF error covariance estimate, our method enables seamless integration of ICP within Kalman filtering, enhancing localization accuracy and robustness. Extensive experiments on the KITTI dataset demonstrate the effectiveness of our approach, showing that it accurately predicts covariance and, when applied to localization using a pre-built map or SLAM, reduces localization errors and improves robustness.

Abstract--While deep learning, including Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), has significantly advanced classification performance, its typical reliance on extensive annotated datasets presents a major obstacle in many practical scenarios where such data is scarce. Vision-language models (VLMs) and transfer learning with pre-trained visual models appear as promising techniques to deal with this problem. This paper proposes a novel zero-shot image classification framework that combines a VLM and a pre-trained visual model within a self-learning cycle. Requiring only the set of class names and no labeled training data, our method utilizes a confidence-based pseudo-labeling strategy to train a lightweight classifier directly on the test data, enabling dynamic adaptation. The VLM identifies high-confidence samples, and the pre-trained visual model enhances their visual representations. These enhanced features then iteratively train the classifier, allowing the system to capture complementary semantic and visual cues without supervision. Notably, our approach avoids VLM fine-tuning and the use of large language models, relying on the visual-only model to reduce the dependence on semantic representation. Experimental evaluations on ten diverse datasets demonstrate that our approach outperforms the baseline zero-shot method. Recent advancements in deep learning have significantly improved image classification performance, making these techniques dominant in current research.

Designing robotic hand morphologies for diverse manipulation tasks requires balancing dexterity, manufacturability, and task-specific functionality. While open-source frameworks and parametric tools support reproducible design, they still rely on expert heuristics and manual tuning. Automated methods using optimization are often compute-intensive, simulation-dependent, and rarely target dexterous hands. Large language models (LLMs), with their broad knowledge of human-object interactions and strong generative capabilities, offer a promising alternative for zero-shot design reasoning. In this paper, we present Lang2Morph, a language-driven pipeline for robotic hand design. It uses LLMs to translate natural-language task descriptions into symbolic structures and OPH-compatible parameters, enabling 3D-printable task-specific morphologies. The pipeline consists of: (i) Morphology Design, which maps tasks into semantic tags, structural grammars, and OPH-compatible parameters; and (ii) Selection and Refinement, which evaluates design candidates based on semantic alignment and size compatibility, and optionally applies LLM-guided refinement when needed. We evaluate Lang2Morph across varied tasks, and results show that our approach can generate diverse, task-relevant morphologies. To our knowledge, this is the first attempt to develop an LLM-based framework for task-conditioned robotic hand design.