Multivariate time series forecasting involves predicting future values based on historical observations. However, existing approaches primarily rely on predefined single-scale patches or lack effective mechanisms for multi-scale feature fusion. These limitations hinder them from fully capturing the complex patterns inherent in time series, leading to constrained performance and insufficient generalizability. To address these challenges, we propose a novel architecture named Adaptive Weighted Mixture of Multi-Scale Expert Transformers (AdaMixT). Specifically, AdaMixT introduces various patches and leverages both General Pre-trained Models (GPM) and Domain-specific Models (DSM) for multi-scale feature extraction. To accommodate the heterogeneity of temporal features, AdaMixT incorporates a gating network that dynamically allocates weights among different experts, enabling more accurate predictions through adaptive multi-scale fusion. Comprehensive experiments on eight widely used benchmarks, including Weather, Traffic, Electricity, ILI, and four ETT datasets, consistently demonstrate the effectiveness of AdaMixT in real-world scenarios.

We study learning of continuous-time inventory dynamics under stochastic demand and quantify when structure helps or hurts forecasting of the bullwhip effect. BULL-ODE compares a fully learned Neural ODE (NODE) that models the entire right-hand side against a physics-informed Universal Differential Equation (UDE) that preserves conservation and order-up-to structure while learning a small residual policy term. Classical supply chain models explain the bullwhip through control/forecasting choices and information sharing, while recent physics-informed and neural differential equation methods blend domain constraints with learned components. It is unclear whether structural bias helps or hinders forecasting under different demand regimes. We address this by using a single-echelon testbed with three demand regimes - AR(1) (autocorrelated), i.i.d. Gaussian, and heavy-tailed lognormal. Training is done on varying fractions of each trajectory, followed by evaluation of multi-step forecasts for inventory I, order rate O, and demand D. Across the structured regimes, UDE consistently generalizes better: with 90% of the training horizon, inventory RMSE drops from 4.92 (NODE) to 0.26 (UDE) under AR(1) and from 5.96 to 0.95 under Gaussian demand. Under heavy-tailed lognormal shocks, the flexibility of NODE is better. These trends persist as train18 ing data shrinks, with NODE exhibiting phase drift in extrapolation while UDE remains stable but underreacts to rare spikes. Our results provide concrete guidance: enforce structure when noise is light-tailed or temporally correlated; relax structure when extreme events dominate. Beyond inventory control, the results offer guidance for hybrid modeling in scientific and engineering systems: enforce known structure when conservation laws and modest noise dominate, and relax structure to capture extremes in settings where rare events drive dynamics.

Test-time scaling (TTS) has been shown to improve the performance of large language models (LLMs) by sampling and aggregating diverse reasoning paths. However, existing research has overlooked a critical issue: selection bias of reasoning strategies during scaling. Specifically, when generating reasoning processes, LLMs tend to follow certain strategies (e.g., algebraic solutions for math problems) while neglecting other valid alternatives (e.g., geometric solutions), resulting in insufficient exploration of the solution space. To further understand the impact of this bias, we present a theoretical analysis that reveals when it undermines the effectiveness of test-time scaling. Motivated by this theoretical insight, we introduce TTS-Uniform, a framework designed to mitigate the selection bias of reasoning strategies. It (i) identifies potential strategies, (ii) uniformly allocates the sampling budget across them, and (iii) filters out unstable strategies prior to aggregation. Experimental results show that TTS-Uniform significantly enhances scaling effectiveness across multiple mainstream LLMs and benchmark datasets. Code is available at https://github.com/zongqianwu/Uniform-TTS. Chain-of-thought (CoT) (Wei et al., 2022; Kojima et al., 2022) enhances the reasoning capabilities of large language models (LLMs) by explicitly unfolding intermediate steps (i.e., reasoning paths) before arriving at the final answer. Building on CoT, test-time scaling (TTS) (Zhang et al., 2025; Ji et al., 2025) further improves performance by sampling and aggregating diverse paths. However, existing TTS research (Wang et al., 2022; Snell et al., 2024) overlooks a critical limitation of CoT, which in turn constrains the effectiveness of scaling. Specifically, when tackling a problem, CoT reasoning tends to follow certain strategies while neglecting other valid alternatives.

Efficient inference is a critical challenge in deep generative modeling, particularly as diffusion models grow in capacity and complexity. While increased complexity often improves accuracy, it raises compute costs, latency, and memory requirements. This work investigates techniques such as pruning, quantization, knowledge distillation, and simplified attention to reduce computational overhead without impacting performance. The study also explores the Mixture of Experts (MoE) approach to further enhance efficiency. These experiments provide insights into optimizing inference for the state-of-the-art Fast Diffusion Transformer (fast-DiT) model.

Generation of Artificial Intelligence (AI) texts in important works has become a common practice that can be used to misuse and abuse AI at various levels. Traditional AI detectors often rely on document-level classification, which struggles to identify AI content in hybrid or slightly edited texts designed to avoid detection, leading to concerns about the model's efficiency, which makes it hard to distinguish between human-written and AI-generated texts. A sentence-level sequence labeling model proposed to detect transitions between human- and AI-generated text, leveraging nuanced linguistic signals overlooked by document-level classifiers. By this method, detecting and segmenting AI and human-written text within a single document at the token-level granularity is achieved. Our model combines the state-of-the-art pre-trained Transformer models, incorporating Neural Networks (NN) and Conditional Random Fields (CRFs). This approach extends the power of transformers to extract semantic and syntactic patterns, and the neural network component to capture enhanced sequence-level representations, thereby improving the boundary predictions by the CRF layer, which enhances sequence recognition and further identification of the partition between Human- and AI-generated texts. The evaluation is performed on two publicly available benchmark datasets containing collaborative human and AI-generated texts. Our experimental comparisons are with zero-shot detectors and the existing state-of-the-art models, along with rigorous ablation studies to justify that this approach, in particular, can accurately detect the spans of AI texts in a completely collaborative text. All our source code and the processed datasets are available in our GitHub repository.

The increasing availability of open Earth Observation (EO) and agricultural datasets holds great potential for supporting sustainable land management. However, their high technical entry barrier limits accessibility for non-expert users. This study presents an open-source conversational assistant that integrates multimodal retrieval and large language models (LLMs) to enable natural language interaction with heterogeneous agricultural and geospatial data. The proposed architecture combines orthophotos, Sentinel-2 vegetation indices, and user-provided documents through retrieval-augmented generation (RAG), allowing the system to flexibly determine whether to rely on multimodal evidence, textual knowledge, or both in formulating an answer. To assess response quality, we adopt an LLM-as-a-judge methodology using Qwen3-32B in a zero-shot, unsupervised setting, applying direct scoring in a multi-dimensional quantitative evaluation framework. Preliminary results show that the system is capable of generating clear, relevant, and context-aware responses to agricultural queries, while remaining reproducible and scalable across geographic regions. The primary contributions of this work include an architecture for fusing multimodal EO and textual knowledge sources, a demonstration of lowering the barrier to access specialized agricultural information through natural language interaction, and an open and reproducible design.

The rapid adoption of large language models (LLMs) in critical domains has spurred extensive research into their security issues. While input manipulation attacks (e.g., prompt injection) have been well studied, Bit-Flip Attacks (BFAs) -- which exploit hardware vulnerabilities to corrupt model parameters and cause severe performance degradation -- have received far less attention. Existing BFA methods suffer from key limitations: they fail to balance performance degradation and output naturalness, making them prone to discovery. In this paper, we introduce SilentStriker, the first stealthy bit-flip attack against LLMs that effectively degrades task performance while maintaining output naturalness. Our core contribution lies in addressing the challenge of designing effective loss functions for LLMs with variable output length and the vast output space. Unlike prior approaches that rely on output perplexity for attack loss formulation, which inevitably degrade output naturalness, we reformulate the attack objective by leveraging key output tokens as targets for suppression, enabling effective joint optimization of attack effectiveness and stealthiness. Additionally, we employ an iterative, progressive search strategy to maximize attack efficacy. Experiments show that SilentStriker significantly outperforms existing baselines, achieving successful attacks without compromising the naturalness of generated text.

Existing table understanding methods face challenges due to complex table structures and intricate logical reasoning. While supervised finetuning (SFT) dominates existing research, reinforcement learning (RL), such as Group Relative Policy Optimization (GRPO), has shown promise but struggled with low initial policy accuracy and coarse rewards in tabular contexts. In this paper, we introduce Table-R1, a three-stage RL framework that enhances multimodal table understanding through: (1) Warm-up that prompts initial perception and reasoning capabilities, (2) Perception Alignment GRPO (PA-GRPO), which employs continuous Tree-Edit-Distance Similarity (TEDS) rewards for recognizing table structures and contents, and (3) Hint-Completion GRPO (HC-GRPO), which utilizes fine-grained rewards of residual steps based on the hint-guided question. Extensive experiments demonstrate that Table-R1 can boost the model's table reasoning performance obviously on both held-in and held-out datasets, outperforming SFT and GRPO largely. Notably, Qwen2-VL-7B with Table-R1 surpasses larger specific table understanding models (e.g., Table-LLaVA 13B), even achieving comparable performance to the closed-source model GPT-4o on held-in datasets, demonstrating the efficacy of each stage of Table-R1 in overcoming initialization bottlenecks and reward sparsity, thereby advancing robust multimodal table understanding.

To mitigate the hallucination problem in large language models, DoLa exploits early exit logits from the same model as a contrastive prior. However, we found that these early exit logits tend to be flat, low in magnitude, and fail to reflect meaningful contrasts. To address this, we propose PruneCD, a novel contrastive decoding method that constructs the amateur model via layer pruning rather than early exit. This design leads to more informative and well-aligned logits, enabling more effective contrastive decoding. Through qualitative and quantitative analyses, we demonstrate that PruneCD consistently improves factuality with minimal inference overhead, offering a robust and practical approach to mitigating hallucinations in LLMs.

In this paper, we introduce a novel weighted co-training approach that is guided by Large Language Models (LLMs). Namely, in our co-training approach, we use LLM labels on unlabeled data as target labels and co-train two encoder-only based networks that train each other over multiple iterations: first, all samples are forwarded through each network and historical estimates of each network's confidence in the LLM label are recorded; second, a dynamic importance weight is derived for each sample according to each network's belief in the quality of the LLM label for that sample; finally, the two networks exchange importance weights with each other -- each network back-propagates all samples weighted with the importance weights coming from its peer network and updates its own parameters. By strategically utilizing LLM-generated guidance, our approach significantly outperforms conventional SSL methods, particularly in settings with abundant unlabeled data. Empirical results show that it achieves state-of-the-art performance on 4 out of 5 benchmark datasets and ranks first among 14 compared methods according to the Friedman test. Our results highlight a new direction in semi-supervised learning -- where LLMs serve as knowledge amplifiers, enabling backbone co-training models to achieve state-of-the-art performance efficiently.