Exponentially growing short video platforms (SVPs) face significant challenges in moderating content detrimental to users' mental health, particularly for minors. The dissemination of such content on SVPs can lead to catastrophic societal consequences. Although substantial efforts have been dedicated to moderating such content, existing methods suffer from critical limitations: (1) Manual review is prone to human bias and incurs high operational costs. (2) Automated methods, though efficient, lack nuanced content understanding, resulting in lower accuracy. (3) Industrial moderation regulations struggle to adapt to rapidly evolving trends due to long update cycles. In this paper, we annotate the first SVP content moderation benchmark with authentic user/reviewer feedback to fill the absence of benchmark in this field. Then we evaluate various methods on the benchmark to verify the existence of the aforementioned limitations. We further propose our common-law content moderation framework named KuaiMod to address these challenges. KuaiMod consists of three components: training data construction, offline adaptation, and online deployment & refinement. Leveraging large vision language model (VLM) and Chain-of-Thought (CoT) reasoning, KuaiMod adequately models video toxicity based on sparse user feedback and fosters dynamic moderation policy with rapid update speed and high accuracy. Offline experiments and large-scale online A/B test demonstrates the superiority of KuaiMod: KuaiMod achieves the best moderation performance on our benchmark. The deployment of KuaiMod reduces the user reporting rate by 20% and its application in video recommendation increases both Daily Active User (DAU) and APP Usage Time (AUT) on several Kuaishou scenarios. We have open-sourced our benchmark at https://kuaimod.github.io.

Machine Unlearning (MUL) is crucial for privacy protection and content regulation, yet recent studies reveal that traces of forgotten information persist in unlearned models, enabling adversaries to resurface removed knowledge. Existing verification methods only confirm whether unlearning was executed, failing to detect such residual information leaks. To address this, we introduce the concept of Robust Unlearning, ensuring models are indistinguishable from retraining and resistant to adversarial recovery. To empirically evaluate whether unlearning techniques meet this security standard, we propose the Unlearning Mapping Attack (UMA), a post-unlearning verification framework that actively probes models for forgotten traces using adversarial queries. Extensive experiments on discriminative and generative tasks show that existing unlearning techniques remain vulnerable, even when passing existing verification metrics. By establishing UMA as a practical verification tool, this study sets a new standard for assessing and enhancing machine unlearning security.

Research on evaluatin g and analyzing large language models (LLMs) has been extensive for high - resource languages such as English, yet their performance in languages such as Persian has received considerably less attention. This paper introduces FarsEval - PKBETS benchmark, a subset of FarsEval project for evaluat ing large language models in Persian. This benchmark consists of 4,000 questions and answers in various formats, including multiple - choice, short - answer, and descriptive responses. It covers a wide range of domains and tasks, including medicine, law, religion, Persian language, encyclopedic knowledge, human preferences, social knowledge, ethics and bias, text generation, and respecting others' rights. This benchmark incorporates linguistic, cultural, and local considera tions relevant to the Persian language and Iran. To ensure the questions are challenging for current LLMs, three models -- Llama3 - 70B, PersianMind, and Dorna -- were evaluated using this benchmark. Their average accuracy was below 50%, meaning they provided full y correct answers to fewer than half of the questions. These results indicate that current language models are still far from being able to solve this benchmark.

In the field of natural language processing (NLP), the capability to infer and reason about causality is increasingly recognized as a critical component of intelligent systems. Despite the recent advancement of large language models (LLMs) (Radford et al., 2019; Devlin et al., 2019; Brown et al., 2020; Zhang et al., 2022; OpenAI, 2023; Ignat et al., 2024, inter alia), a key question still remains: Can these models understand and reason about causality? This is a critical skill before we can trust AI agents to be integrated into decision-making systems. Moreover, even if LLMs succeed at some extent of reasoning, they still lack transparency of how their decisions are made, forming a strong need for interpretabil-ity (Luo and Specia, 2024; Räuker et al., 2023; Zou et al., 2023). T o bridge the gap, this thesis explores various facets of causal reasoning in LLMs. W e present a series of studies that collectively advance the knowledge of how well these models perform causal reasoning (Part I), how their decisions are made (Part II), how causality among learning variables influences NLP tasks (Part III), and how causality and NLP can together analyze social problems (Part IV). Below we introduce an overview of the four parts and their corresponding chapters.

Atom probe tomography (APT) is a burgeoning characterization technique that provides compositional mapping of materials in three-dimensions at near-atomic scale. Since its significant expansion in the past 30 years, we estimate that one million APT datasets have been collected, each containing millions to billions of individual ions. Their analysis and the extraction of microstructural information has largely relied upon individual users whose varied level of expertise causes clear and documented bias. Current practices hinder efficient data processing, and make challenging standardization and the deployment of data analysis workflows that would be compliant with FAIR data principles. Over the past decade, building upon the long-standing expertise of the APT community in the development of advanced data processing or data mining techniques, there has been a surge of novel machine learning (ML) approaches aiming for user-independence, and that are efficient, reproducible, and robust from a statistics perspective. Here, we provide a snapshot review of this rapidly evolving field. We begin with a brief introduction to APT and the nature of the APT data. This is followed by an overview of relevant ML algorithms and a comprehensive review of their applications to APT. We also discuss how ML can enable discoveries beyond human capability, offering new insights into the mechanisms within materials. Finally, we provide guidance for future directions in this domain.

The rapid development of video surveillance systems for object detection, tracking, activity recognition, and anomaly detection has revolutionized our day-to-day lives while setting alarms for privacy concerns. It isn't easy to strike a balance between visual privacy and action recognition performance in most computer vision models. Is it possible to safeguard privacy without sacrificing performance? It poses a formidable challenge, as even minor privacy enhancements can lead to substantial performance degradation. To address this challenge, we propose a privacy-preserving image anonymization technique that optimizes the anonymizer using penalties from the utility branch, ensuring improved action recognition performance while minimally affecting privacy leakage. This approach addresses the trade-off between minimizing privacy leakage and maintaining high action performance. The proposed approach is primarily designed to align with the regulatory standards of the EU AI Act and GDPR, ensuring the protection of personally identifiable information while maintaining action performance. To the best of our knowledge, we are the first to introduce a feature-based penalty scheme that exclusively controls the action features, allowing freedom to anonymize private attributes. Extensive experiments were conducted to validate the effectiveness of the proposed method. The results demonstrate that applying a penalty to anonymizer from utility branch enhances action performance while maintaining nearly consistent privacy leakage across different penalty settings.

Large Language Models (LLMs) have transformed natural language processing, but concerns have emerged about their susceptibility to ideological manipulation, particularly in politically sensitive areas. Prior work has focused on binary Left-Right LLM biases, using explicit prompts and fine-tuning on political QA datasets. In this work, we move beyond this binary approach to explore the extent to which LLMs can be influenced across a spectrum of political ideologies, from Progressive-Left to Conservative-Right. We introduce a novel multi-task dataset designed to reflect diverse ideological positions through tasks such as ideological QA, statement ranking, manifesto cloze completion, and Congress bill comprehension. By fine-tuning three LLMs-Phi-2, Mistral, and Llama-3-on this dataset, we evaluate their capacity to adopt and express these nuanced ideologies. Our findings indicate that fine-tuning significantly enhances nuanced ideological alignment, while explicit prompts provide only minor refinements. This highlights the models' susceptibility to subtle ideological manipulation, suggesting a need for more robust safeguards to mitigate these risks.

Large models such as Large Language Models (LLMs) and Vision Language Models (VLMs) have transformed artificial intelligence, powering applications in natural language processing, computer vision, and multimodal learning. However, fully fine-tuning these models remains expensive, requiring extensive computational resources, memory, and task-specific data. Parameter-Efficient Fine-Tuning (PEFT) has emerged as a promising solution that allows adapting large models to downstream tasks by updating only a small portion of parameters. This survey presents a comprehensive overview of PEFT techniques, focusing on their motivations, design principles, and effectiveness. We begin by analyzing the resource and accessibility challenges posed by traditional fine-tuning and highlight key issues, such as overfitting, catastrophic forgetting, and parameter inefficiency. We then introduce a structured taxonomy of PEFT methods -- grouped into additive, selective, reparameterized, hybrid, and unified frameworks -- and systematically compare their mechanisms and trade-offs. Beyond taxonomy, we explore the impact of PEFT across diverse domains, including language, vision, and generative modeling, showing how these techniques offer strong performance with lower resource costs. We also discuss important open challenges in scalability, interpretability, and robustness, and suggest future directions such as federated learning, domain adaptation, and theoretical grounding. Our goal is to provide a unified understanding of PEFT and its growing role in enabling practical, efficient, and sustainable use of large models.

Current AI models often fail to account for local context and language, given the predominance of English and Western internet content in their training data. This hinders the global relevance, usefulness, and safety of these models as they gain more users around the globe. Amplify Initiative, a data platform and methodology, leverages expert communities to collect diverse, high-quality data to address the limitations of these models. The platform is designed to enable co-creation of datasets, provide access to high-quality multilingual datasets, and offer recognition to data authors. This paper presents the approach to co-creating datasets with domain experts (e.g., health workers, teachers) through a pilot conducted in Sub-Saharan Africa (Ghana, Kenya, Malawi, Nigeria, and Uganda). In partnership with local researchers situated in these countries, the pilot demonstrated an end-to-end approach to co-creating data with 155 experts in sensitive domains (e.g., physicians, bankers, anthropologists, human and civil rights advocates). This approach, implemented with an Android app, resulted in an annotated dataset of 8,091 adversarial queries in seven languages (e.g., Luganda, Swahili, Chichewa), capturing nuanced and contextual information related to key themes such as misinformation and public interest topics. This dataset in turn can be used to evaluate models for their safety and cultural relevance within the context of these languages.

This work defends the 'Whole Hog Thesis': sophisticated Large Language Models (LLMs) like ChatGPT are full-blown linguistic and cognitive agents, possessing understanding, beliefs, desires, knowledge, and intentions. We argue against prevailing methodologies in AI philosophy, rejecting starting points based on low-level computational details ('Just an X' fallacy) or pre-existing theories of mind. Instead, we advocate starting with simple, high-level observations of LLM behavior (e.g., answering questions, making suggestions) -- defending this data against charges of metaphor, loose talk, or pretense. From these observations, we employ 'Holistic Network Assumptions' -- plausible connections between mental capacities (e.g., answering implies knowledge, knowledge implies belief, action implies intention) -- to argue for the full suite of cognitive states. We systematically rebut objections based on LLM failures (hallucinations, planning/reasoning errors), arguing these don't preclude agency, often mirroring human fallibility. We address numerous 'Games of Lacks', arguing that LLMs do not lack purported necessary conditions for cognition (e.g., semantic grounding, embodiment, justification, intrinsic intentionality) or that these conditions are not truly necessary, often relying on anti-discriminatory arguments comparing LLMs to diverse human capacities. Our approach is evidential, not functionalist, and deliberately excludes consciousness. We conclude by speculating on the possibility of LLMs possessing 'alien' contents beyond human conceptual schemes.