This paper introduces an Artificial Intelligence (AI) foundation model for time series in engineering applications, where causal operations are required for real-time monitoring and control. Since engineering time series are governed by physical, rather than linguistic, laws, large-language-model-based AI foundation models may be ineffective or inefficient. Building on the classical innovations representation theory of Wiener, Kallianpur, and Rosenblatt, we propose Time Series GPT (TS-GPT) -- an innovations-representation-based Generative Pre-trained Transformer for engineering monitoring and control. As an example of foundation model adaptation, we consider Probabilistic Generative Forecasting, which produces future time series samples from conditional probability distributions given past realizations. We demonstrate the effectiveness of TS-GPT in forecasting real-time locational marginal prices using historical data from U.S. independent system operators.

Standard discrete diffusion models treat all unobserved states identically by mapping them to an absorbing [MASK] token. This creates an 'information void' where semantic information that could be inferred from unmasked tokens is lost between denoising steps. We introduce Continuously Augmented Discrete Diffusion (CADD), a framework that augments the discrete state space with a paired diffusion in a continuous latent space. This yields graded, gradually corrupted states in which masked tokens are represented by noisy yet informative latent vectors rather than collapsed 'information voids'. At each reverse step, CADD may leverage the continuous latent as a semantic hint to guide discrete denoising. The design is clean and compatible with existing discrete diffusion training. At sampling time, the strength and choice of estimator for the continuous latent vector enables a controlled trade-off between mode-coverage (generating diverse outputs) and mode-seeking (generating contextually precise outputs) behaviors. Empirically, we demonstrate CADD improves generative quality over mask-based diffusion across text generation, image synthesis, and code modeling, with consistent gains on both qualitative and quantitative metrics against strong discrete baselines.

We study the problem of determining whether a piece of text has been authored by a human or by a large language model (LLM). Existing state of the art logits-based detectors make use of statistics derived from the log-probability of the observed text evaluated using the distribution function of a given source LLM. However, relying solely on log probabilities can be sub-optimal. In response, we introduce AdaDetectGPT -- a novel classifier that adaptively learns a witness function from training data to enhance the performance of logits-based detectors. We provide statistical guarantees on its true positive rate, false positive rate, true negative rate and false negative rate. Extensive numerical studies show AdaDetectGPT nearly uniformly improves the state-of-the-art method in various combination of datasets and LLMs, and the improvement can reach up to 58%. A python implementation of our method is available at https://github.com/Mamba413/AdaDetectGPT.

Linking textual values in tabular data to their corresponding entities in a Knowledge Base is a core task across a variety of data integration and enrichment applications. Although Large Language Models (LLMs) have shown State-of-The-Art performance in Entity Linking (EL) tasks, their deployment in real-world scenarios requires not only accurate predictions but also reliable uncertainty estimates, which require resource-demanding multi-shot inference, posing serious limits to their actual applicability. As a more efficient alternative, we investigate a self-supervised approach for estimating uncertainty from single-shot LLM outputs using token-level features, reducing the need for multiple generations. Evaluation is performed on an EL task on tabular data across multiple LLMs, showing that the resulting uncertainty estimates are highly effective in detecting low-accuracy outputs. This is achieved at a fraction of the computational cost, ultimately supporting a cost-effective integration of uncertainty measures into LLM-based EL workflows. The method offers a practical way to incorporate uncertainty estimation into EL workflows with limited computational overhead.

This study addresses the reliability of automatic summarization in high-risk scenarios and proposes a large language model framework that integrates uncertainty quantification and risk-aware mechanisms. Starting from the demands of information overload and high-risk decision-making, a conditional generation-based summarization model is constructed, and Bayesian inference is introduced during generation to model uncertainty in the parameter space, which helps avoid overconfident predictions. The uncertainty level of the generated content is measured using predictive distribution entropy, and a joint optimization of entropy regularization and risk-aware loss is applied to ensure that key information is preserved and risk attributes are explicitly expressed during information compression. On this basis, the model incorporates risk scoring and regulation modules, allowing summaries to cover the core content accurately while enhancing trustworthiness through explicit risk-level prompts. Comparative experiments and sensitivity analyses verify that the proposed method significantly improves the robustness and reliability of summarization in high-risk applications while maintaining fluency and semantic integrity. This research provides a systematic solution for trustworthy summarization and demonstrates both scalability and practical value at the methodological level.

Large language models (LLMs) are increasingly used in clinical settings, yet their performance in veterinary medicine remains underexplored. We evaluated three commercially available veterinary-focused LLM summarization tools (Product 1 [Hachiko] and Products 2 and 3) on a standardized dataset of veterinary oncology records. Using a rubric-guided LLM-as-a-judge framework, summaries were scored across five domains: Factual Accuracy, Completeness, Chronological Order, Clinical Relevance, and Organization. Product 1 achieved the highest overall performance, with a median average score of 4.61 (IQR: 0.73), compared to 2.55 (IQR: 0.78) for Product 2 and 2.45 (IQR: 0.92) for Product 3. It also received perfect median scores in Factual Accuracy and Chronological Order. To assess the internal consistency of the grading framework itself, we repeated the evaluation across three independent runs. The LLM grader demonstrated high reproducibility, with Average Score standard deviations of 0.015 (Product 1), 0.088 (Product 2), and 0.034 (Product 3). These findings highlight the importance of veterinary-specific commercial LLM tools and demonstrate that LLM-as-a-judge evaluation is a scalable and reproducible method for assessing clinical NLP summarization in veterinary medicine.

As Large Language Models (LLMs) transition from static tools to autonomous agents, traditional evaluation benchmarks that measure performance on downstream tasks are becoming insufficient. These methods fail to capture the emergent social and cognitive dynamics that arise when agents communicate, persuade, and collaborate in interactive environments. To address this gap, we introduce a novel evaluation framework that uses multi-agent debate as a controlled "social laboratory" to discover and quantify these behaviors. In our framework, LLM-based agents, instantiated with distinct personas and incentives, deliberate on a wide range of challenging topics under the supervision of an LLM moderator. Our analysis, enabled by a new suite of psychometric and semantic metrics, reveals several key findings. Across hundreds of debates, we uncover a powerful and robust emergent tendency for agents to seek consensus, consistently reaching high semantic agreement (μ > 0.88) even without explicit instruction and across sensitive topics. We show that assigned personas induce stable, measurable psychometric profiles, particularly in cognitive effort, and that the moderators persona can significantly alter debate outcomes by structuring the environment, a key finding for external AI alignment. This work provides a blueprint for a new class of dynamic, psychometrically grounded evaluation protocols designed for the agentic setting, offering a crucial methodology for understanding and shaping the social behaviors of the next generation of AI agents. We have released the code and results at https://github.com/znreza/multi-agent-LLM-eval-for-debate.

We study the problem of zero-shot link prediction on knowledge graphs (KGs), which requires models to generalize over novel entities and novel relations. Knowledge graph foundation models (KGFMs) address this task by enforcing equivariance over both nodes and relations, learning from structural properties of nodes and relations, which are then transferable to novel graphs with similar structural properties. However, the conventional notion of deterministic equivariance imposes inherent limits on the expressive power of KGFMs, preventing them from distinguishing structurally similar but semantically distinct relations. To overcome this limitation, we introduce probabilistic node-relation equivariance, which preserves equivariance in distribution while incorporating a principled randomization to break symmetries during inference. Building on this principle, we present Flock, a KGFM that iteratively samples random walks, encodes them into sequences via a recording protocol, embeds them with a sequence model, and aggregates representations of nodes and relations via learned pooling. Crucially, Flock respects probabilistic node-relation equivariance and is a universal approximator for isomorphism-invariant link-level functions over KGs. Empirically, Flock perfectly solves our new diagnostic dataset Petals where current KGFMs fail, and achieves state-of-the-art performances on entity- and relation prediction tasks on 54 KGs from diverse domains.

Humans intuitively perceive complex social signals in visual scenes, yet it remains unclear whether state-of-the-art AI models encode the same similarity structure. We study (Q1) whether modern video and language models capture human-perceived similarity in social videos, and (Q2) how to instill this structure into models using human behavioral data. To address this, we introduce a new benchmark of over 49,000 odd-one-out similarity judgments on 250 three-second video clips of social interactions, and discover a modality gap: despite the task being visual, caption-based language embeddings align better with human similarity than any pretrained video model. We close this gap by fine-tuning a TimeSformer video model on these human judgments with our novel hybrid triplet-RSA objective using low-rank adaptation (LoRA), aligning pairwise distances to human similarity. This fine-tuning protocol yields significantly improved alignment with human perceptions on held-out videos in terms of both explained variance and odd-one-out triplet accuracy. Variance partitioning shows that the fine-tuned video model increases shared variance with language embeddings and explains additional unique variance not captured by the language model. Finally, we test transfer via linear probes and find that human-similarity fine-tuning strengthens the encoding of social-affective attributes (intimacy, valence, dominance, communication) relative to the pretrained baseline. Overall, our findings highlight a gap in pretrained video models' social recognition and demonstrate that behavior-guided fine-tuning shapes video representations toward human social perception.

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a key method for improving Large Language Models' reasoning capabilities, yet recent evidence suggests it may paradoxically shrink the reasoning boundary rather than expand it. This paper investigates the shrinkage issue of RLVR by analyzing its learning dynamics and reveals two critical phenomena that explain this failure. First, we expose negative interference in RLVR, where learning to solve certain training problems actively reduces the likelihood of correct solutions for others, leading to the decline of Pass@$k$ performance, or the probability of generating a correct solution within $k$ attempts. Second, we uncover the winner-take-all phenomenon: RLVR disproportionately reinforces problems with high likelihood, correct solutions, under the base model, while suppressing other initially low-likelihood ones. Through extensive theoretical and empirical analysis on multiple mathematical reasoning benchmarks, we show that this effect arises from the inherent on-policy sampling in standard RL objectives, causing the model to converge toward narrow solution strategies. Based on these insights, we propose a simple yet effective data curation algorithm that focuses RLVR learning on low-likelihood problems, achieving notable improvement in Pass@$k$ performance. Our code is available at https://github.com/mail-research/SELF-llm-interference.