For example, applications applications have become increasingly popular. In this paper, such as book summarization [12-14], document classification through trace-based experiments, we found that the existing [15, 16], and coding assistance [17] require a longer or method for long sequences results in a high Time-To-unlimited sequence length to fully understand the extended First-Token (TTFT) due to sequential chunk processing, long context. Some long-sequence applications, such as coding assistance, Time-Between-Tokens (TBT) from batching long-sequence require short response time (e.g., in seconds). However, prefills and decodes, and low throughput due to constrained through experimental measurements, we made Observation key-value cache (KVC) for long sequences. To address these (O): issues, we propose two Sequence-Parallelism (SP) architectures O1. The existing serving system that handles long sequences, for both tensor parallelism (TP) and non-TP. However, Sarathi-Serve [18], generates long Time-To-First-Token SP introduces two challenges: 1) network communication (TTFT) (in minutes) due to sequential chunk processing, high and computation become performance bottlenecks; 2) the Time-Between-Token (TBT) (e.g., 6 seconds) from batching latter two issues above are mitigated but not resolved, and long-sequence prefills and decodes, and low throughput due SP's resultant KV value distribution across GPUs still requires to small batch size caused by constrained KV cache size and communication for decode, increasing TBT.

Recently, models for user representation learning have been widely applied in click-through-rate (CTR) and conversion-rate (CVR) prediction. Usually, the model learns a universal user representation as the input for subsequent scenario-specific models. However, in numerous industrial applications (e.g., recommendation and marketing), the business always operates such applications as various online activities among different user segmentation. These segmentation are always created by domain experts. Due to the difference in user distribution (i.e., user segmentation) and business objectives in subsequent tasks, learning solely on universal representation may lead to detrimental effects on both model performance and robustness. In this paper, we propose a novel learning framework that can first learn general universal user representation through information bottleneck. Then, merge and learn a segmentation-specific or a task-specific representation through neural interaction. We design the interactive learning process by leveraging a bipartite graph architecture to model the representation learning and merging between contextual clusters and each user segmentation. Our proposed method is evaluated in two open-source benchmarks, two offline business datasets, and deployed on two online marketing applications to predict users' CVR. The results demonstrate that our method can achieve superior performance and surpass the baseline methods.

Reinforcement learning methods have been used to optimize long-term user engagement in recommendation systems. However, existing reinforcement learning-based recommendation systems do not fully exploit the relevance of individual user behavior across different platforms. One potential solution is to aggregate data from various platforms in a centralized location and use the aggregated data for training. However, this approach raises economic and legal concerns, including increased communication costs and potential threats to user privacy. To address these challenges, we propose \textbf{FedSlate}, a federated reinforcement learning recommendation algorithm that effectively utilizes information that is prohibited from being shared at a legal level. We employ the SlateQ algorithm to assist FedSlate in learning users' long-term behavior and evaluating the value of recommended content. We extend the existing application scope of recommendation systems from single-user single-platform to single-user multi-platform and address cross-platform learning challenges by introducing federated learning. We use RecSim to construct a simulation environment for evaluating FedSlate and compare its performance with state-of-the-art benchmark recommendation models. Experimental results demonstrate the superior effects of FedSlate over baseline methods in various environmental settings, and FedSlate facilitates the learning of recommendation strategies in scenarios where baseline methods are completely inapplicable. Code is available at \textit{https://github.com/TianYaDY/FedSlate}.

DPO is an effective preference optimization algorithm. However, the DPO-tuned models tend to overfit on the dispreferred samples, manifested as overly long generations lacking diversity. While recent regularization approaches have endeavored to alleviate this issue by modifying the objective function, they achieved that at the cost of alignment performance degradation. In this paper, we innovatively incorporate regularization from the perspective of weight updating to curb alignment overfitting. Through the pilot experiment, we discovered that there exists a positive correlation between overfitting and the hyperspherical energy fluctuation. Hence, we introduce orthogonal finetuning for DPO via a weight-Rotated Preference Optimization (RoPO) method, which merely conducts rotational and magnitude-stretching updates on the weight parameters to maintain the hyperspherical energy invariant, thereby preserving the knowledge encoded in the angle between neurons. Extensive experiments demonstrate that our model aligns perfectly with human preferences while retaining the original expressive capacity using only 0.0086% of the trainable parameters, suggesting an effective regularization against overfitting. Specifically, RoPO outperforms DPO by up to 10 points on MT-Bench and by up to 2.8 points on AlpacaEval 2, while enhancing the generation diversity by an average of 6 points.

In different NLP tasks, detecting harmful content is crucial for online environments, especially with the growing influence of social media. However, previous research has two main issues: 1) a lack of data in low-resource settings, and 2) inconsistent definitions and criteria for judging harmful content, requiring classification models to be robust to spurious features and diverse. We propose Toxicraft, a novel framework for synthesizing datasets of harmful information to address these weaknesses. With only a small amount of seed data, our framework can generate a wide variety of synthetic, yet remarkably realistic, examples of toxic information. Experimentation across various datasets showcases a notable enhancement in detection model robustness and adaptability, surpassing or close to the gold labels. We release the generated data at Github upon acceptance.

Current video summarization methods largely rely on transformer-based architectures, which, due to their quadratic complexity, require substantial computational resources. In this work, we address these inefficiencies by enhancing the Direct-to-Summarize Network (DSNet) with more resource-efficient token mixing mechanisms. We show that replacing traditional attention with alternatives like Fourier, Wavelet transforms, and Nystr\"omformer improves efficiency and performance. Furthermore, we explore various pooling strategies within the Regional Proposal Network, including ROI pooling, Fast Fourier Transform pooling, and flat pooling. Our experimental results on TVSum and SumMe datasets demonstrate that these modifications significantly reduce computational costs while maintaining competitive summarization performance. Thus, our work offers a more scalable solution for video summarization tasks.

Despite decades of study, many unknowns exist about the mechanisms governing human locomotion. Current models and motor control theories can only partially capture the phenomenon. This may be a major cause of the reduced efficacy of lower limb rehabilitation therapies. Recently, it has been proposed that human locomotion can be planned in the task-space by taking advantage of the gravitational pull acting on the Centre of Mass (CoM) by modelling the attractor dynamics. The model proposed represents the CoM transversal trajectory as a harmonic oscillator propagating on the attractor manifold. However, the vertical trajectory of the CoM, controlled through ankle strategies, has not been accurately captured yet. Research Questions: Is it possible to improve the model accuracy by introducing a mathematical model of the ankle strategies by coordinating the heel-strike and toe-off strategies with the CoM movement? Our solution consists of closed-form equations that plan human-like trajectories for the CoM, the foot swing, and the ankle strategies. We have tested our model by extracting the biomechanics data and postural during locomotion from the motion capture trajectories of 12 healthy subjects at 3 self-selected speeds to generate a virtual subject using our model. Our virtual subject has been based on the average of the collected data. The model output shows our virtual subject has walking trajectories that have their features consistent with our motion capture data. Additionally, it emerged from the data analysis that our model regulates the stance phase of the foot as humans do. The model proves that locomotion can be modelled as an attractor dynamics, proving the existence of a nonlinear map that our nervous system learns. It can support a deeper investigation of locomotion motor control, potentially improving locomotion rehabilitation and assistive technologies.

Accurately estimating data in sensor-less areas is crucial for understanding system dynamics, such as traffic state estimation and environmental monitoring. This study addresses challenges posed by sparse sensor deployment and unreliable data by framing the problem as a spatiotemporal kriging task and proposing a novel graph transformer model, Kriformer. This model estimates data at locations without sensors by mining spatial and temporal correlations, even with limited resources. Kriformer utilizes transformer architecture to enhance the model's perceptual range and solve edge information aggregation challenges, capturing spatiotemporal information effectively. A carefully constructed positional encoding module embeds the spatiotemporal features of nodes, while a sophisticated spatiotemporal attention mechanism enhances estimation accuracy. The multi-head spatial interaction attention module captures subtle spatial relationships between observed and unobserved locations. During training, a random masking strategy prompts the model to learn with partial information loss, allowing the spatiotemporal embedding and multi-head attention mechanisms to synergistically capture correlations among locations. Experimental results show that Kriformer excels in representation learning for unobserved locations, validated on two real-world traffic speed datasets, demonstrating its effectiveness in spatiotemporal kriging tasks.

The integration of Artificial Intelligence (AI) and the Internet of Medical Things (IoMT) in healthcare, through Machine Learning (ML) and Deep Learning (DL) techniques, has advanced the prediction and diagnosis of chronic diseases. AI-driven models such as XGBoost, Random Forest, CNNs, and LSTM RNNs have achieved over 98\% accuracy in predicting heart disease, chronic kidney disease (CKD), Alzheimer's disease, and lung cancer, using datasets from platforms like Kaggle, UCI, private institutions, and real-time IoMT sources. However, challenges persist due to variations in data quality, patient demographics, and formats from different hospitals and research sources. The incorporation of IoMT data, which is vast and heterogeneous, adds complexities in ensuring interoperability and security to protect patient privacy. AI models often struggle with overfitting, performing well in controlled environments but less effectively in real-world clinical settings. Moreover, multi-morbidity scenarios especially for rare diseases like dementia, stroke, and cancers remain insufficiently addressed. Future research should focus on data standardization and advanced preprocessing techniques to improve data quality and interoperability. Transfer learning and ensemble methods are crucial for improving model generalizability across clinical settings. Additionally, the exploration of disease interactions and the development of predictive models for chronic illness intersections is needed. Creating standardized frameworks and open-source tools for integrating federated learning, blockchain, and differential privacy into IoMT systems will also ensure robust data privacy and security.

Biological and artificial neural networks develop internal representations that enable them to perform complex tasks. In artificial networks, the effectiveness of these models relies on their ability to build task specific representation, a process influenced by interactions among datasets, architectures, initialization strategies, and optimization algorithms. Prior studies highlight that different initializations can place networks in either a lazy regime, where representations remain static, or a rich/feature learning regime, where representations evolve dynamically. Here, we examine how initialization influences learning dynamics in deep linear neural networks, deriving exact solutions for lambda-balanced initializations-defined by the relative scale of weights across layers. These solutions capture the evolution of representations and the Neural Tangent Kernel across the spectrum from the rich to the lazy regimes. Our findings deepen the theoretical understanding of the impact of weight initialization on learning regimes, with implications for continual learning, reversal learning, and transfer learning, relevant to both neuroscience and practical applications.