Recent AI agents, such as ChatGPT and LLaMA, primarily rely on instruction tuning and reinforcement learning to calibrate the output of large language models (LLMs) with human intentions, ensuring the outputs are harmless and helpful. Existing methods heavily depend on the manual annotation of high-quality positive samples, while contending with issues such as noisy labels and minimal distinctions between preferred and dispreferred response data. However, readily available toxic samples with clear safety distinctions are often filtered out, removing valuable negative references that could aid LLMs in safety alignment. In response, we propose Positive–Toxic Self-Alignment (PT-ALIGN), a novel safety self-alignment approach that minimizes human supervision by automatically refining positive and toxic samples and performing fine-grained dual instruction tuning. Positive samples are harmless responses, while toxic samples deliberately contain extremely harmful content, serving as a new supervisory signal. Specifically, we utilize LLM itself to iteratively generate and refine training instances by only exploring fewer than 50 human annotations. We then employ two losses, i.e., maximum likelihood estimation (MLE) and fine-grained unlikelihood training (UT), to jointly learn to enhance the LLM’s safety. The MLE loss encourages an LLM to maximize the generation of harmless content based on positive samples. Conversely, the fine-grained UT loss guides the LLM to minimize the output of harmful words based on toxic samples at the token-level, thereby guiding the model to decouple safety from effectiveness, directing it toward safer fine-tuning objectives, and increasing the likelihood of generating helpful and reliable content. Experiments on 9 popular open-source LLMs demonstrate the effectiveness of our PT-ALIGN for safety alignment, while maintaining comparable levels of helpfulness and usefulness.

The development of Reasoning Large Language Models (RLLMs) has significantly improved multi-step reasoning capabilities, but it has also made hallucination problems more frequent and harder to eliminate. While existing approaches address hallucination through external knowledge integration, model parameter analysis, or self-verification mechanisms, they fail to provide a comprehensive insight into how hallucinations **emerge** and **evolve** throughout the reasoning chain. In this work, we investigate hallucination causality under constrained knowledge domains by auditing the Chain-of-Thought (CoT) trajectory and assessing the model's cognitive confidence in potentially erroneous or biased claims. Analysis reveals that in long-CoT settings, RLLMs may iteratively reinforce biases and errors through flawed reflective processes, ultimately inducing hallucinated reasoning paths. Counterintuitively, even with interventions at hallucination origins, reasoning chains display pronounced ''chain disloyalty'', resisting correction and sustaining flawed trajectories. We further point out that existing hallucination detection methods are *less reliable and interpretable than previously assumed*, especially in complex multi-step reasoning contexts. Unlike circuit tracing that requires access to model parameters, our auditing **enables more interpretable long-chain hallucination attribution in black-box settings**, demonstrating stronger generalizability and practical utility. Our code is available at [this link](https://github.com/Winnie-Lian/AHa_Meta_Cognitive).

Mobile Internet‐based intelligent media has become a popular academic topic. This study uses the CiteSpace visualisation tool and Scientific Citation Index Expanded database to comb the existing research in the field of intelligent media from a quantitative perspective. A total of 7248 English papers were published on the topic of “intelligent media” from 2012 to 2021, and 145 highly cited papers refined were analysed. Scientific knowledge graphs were analysed from six dimensions: annual publication quantity, country of publication, institution of publication, author, keywords, and cited references. In the last 10 years, the research literature on intelligent media has been found to increase annually. Presently, the People’s Republic of China and the United States of America have a high proportion of documents in this field. Chinese universities and institutions have achieved significantly in terms of the quantity and quality of documents. From the perspective of the whole intelligent media discipline, the high‐yield author group has not been formed, and there is minimal cooperation amongst authors. Popular intelligent media topics include film, social media, machine learning, swarming motility, data mining, and artificial intelligence. Subject words of the main research directions are event recognition, fake news, Cable News Network model, reconfigurable intelligent surface, comprehensive survey, microblog message, strain sensor, and traffic event. Combined with popular topics and time zone maps, the future research frontier in the field of smart media is identified.

We study the min-cost seed selection problem in online social networks for viral marketing, where the goal is to select a set of seed nodes with the minimum total cost such that the expected number of influenced nodes in the network exceeds a predefined threshold. We propose several algorithms that outperform the previous studies both on the theoretical approximation ratio and on the experimental performance. In the case where the nodes have heterogeneous costs, our algorithms are the first bi-criteria approximation algorithms with polynomial running time and provable approximation ratio. In the case where the users have uniform costs, our algorithms achieve logarithmic approximation ratio and provable time complexity which is smaller than that of the existing algorithms in orders of magnitude. We conduct extensive experiments using real social networks. The experimental results show that, our algorithms significantly outperform the existing algorithms both on the total cost and on the running time, and also scale well to billion-scale networks.

One main challenge in developing a system for visual surveillance event detection is the annotation of target events in the training data. By making use of the assumption that events with security interest are often rare compared to regular behaviours, this paper presents a novel approach by using Kullback-Leibler (KL) divergence for rare event detection in a weakly supervised learning setting, where only clip-level annotation is available. It will be shown that this approach outperforms state-of-the-art methods on a popular real-world dataset, while preserving real time performance.

Due to the popularity of security cameras in public places, it is of interest to design an intelligent system that can efficiently detect events automatically. This paper proposes a novel algorithm for multiperson event detection. To ensure greater than real-time performance, features are extracted directly from compressed MPEG video. A novel histogram-based feature descriptor that captures the angles between extracted particle trajectories is proposed, which allows us to capture motion patterns for multiperson events in the video. To alleviate the need for fine-grained annotation, we propose the use of labeled latent Dirichlet allocation, a weakly supervised method that allows the use of coarse temporal annotations, which are much simpler to obtain. This novel system is able to run at ~10 times real time, while preserving state-of-the-art detection performance for multiperson events on a 100-h real-world surveillance data set (TRECVid surveillance event detection).

The ability to detect unusual events in surviellance footage as they happen is a highly desireable feature for a surveillance system. However, this problem remains challenging in crowded scenes due to occlusions and the clustering of people. In this paper, we propose using the Distributed Behavior Model (DBM), which has been widely used in computer graphics, for video event detection. Our approach does not rely on object tracking, and is robust to camera movements. We use sparse coding for classification, and test our approach on various datasets. Our proposed approach outperforms a state-of-the-art work which uses the social force model and Latent Dirichlet Allocation.

Modelling activities in crowded scenes is very challenging as object tracking is not robust in complicated scenes and optical flow does not capture long range motion. We propose a novel approach to analyse activities in crowded scenesusing a "bag of particle trajectories". Particle trajectoriesare extracted from foreground regions within short video clips using particle video, which estimates long rangemotion in contrast to optical flow which is only concerned with inter-frame motion. Our applications include temporal video segmentation and anomaly detection, and we perform our evaluation on several real-world datasets containing complicated scenes. We show that our approaches achieve state-of-the-art performance for both tasks.