We introduce a benchmark framework developed by and for the scientific community to evaluate, monitor and steer large language model development in fundamental physics. Building on philosophical concepts of scientific understanding and creativity, we develop a scoring system in which each question is scored by an expert for its correctness, difficulty, and surprise. The questions are of three forms: (i) multiple-choice questions for conceptual understanding, (ii) analytical problems requiring mathematical derivation, and (iii) openended tasks requiring complex problem solving. Our current dataset contains diverse set of examples, including a machine learning challenge to classify high-energy physics events, such as the four top quark signal. To ensure continued relevance, we propose a living benchmark, where physicists contribute questions, for instance alongside new publications. We invite contributions via: http://www.physicsbenchmarks.org/. We hope that this benchmark will enable a targeted AI development that can make a meaningful contribution to fundamental physics research.

As large language models (LLMs) rapidly advance, their role in code generation has expanded significantly. While this offers streamlined development, it also creates concerns in areas like education and job interviews. Consequently, developing robust systems to detect AI-generated code is imperative to maintain academic integrity and ensure fairness in hiring processes. In this study, we introduce MultiAIGCD, a dataset for AI-generated code detection for Python, Java, and Go. From the CodeNet dataset's problem definitions and human-authored codes, we generate several code samples in Java, Python, and Go with six different LLMs and three different prompts. This generation process covered three key usage scenarios: (i) generating code from problem descriptions, (ii) fixing runtime errors in human-written code, and (iii) correcting incorrect outputs. Overall, MultiAIGCD consists of 121,271 AI-generated and 32,148 human-written code snippets. We also benchmark three state-of-the-art AI-generated code detection models and assess their performance in various test scenarios such as cross-model and cross-language. We share our dataset and codes to support research in this field.

Audio-visual multi-task incremental learning aims to continuously learn from multiple audio-visual tasks without the need for joint training on all tasks. The challenge of the problem is how to preserve the old task knowledge while facilitating the learning of new task with previous experiences. T o address these challenges, we introduce a three-stage Progressive Homeostatic and Plastic audio-visual prompt (PHP) method. In the shallow phase, we design the task-shared modality aggregating adapter to foster cross-task and cross-modal audio-visual representation learning to enhance shared understanding between tasks. In the middle phase, we propose the task-specific modality-shared dynamic generating adapter, which constructs prompts that are tailored to individual tasks while remaining general across modalities, which balances the model's ability to retain knowledge against forgetting with its potential for versatile multi-task transferability. In the deep phase, we introduce the task-specific modality-independent prompts to further refine the understand ability by targeting individual information for each task and modality. By incorporating these three phases, PHP retains task-specific prompts while adapting shared parameters for new tasks to effectively balance knowledge sharing and specificity. Our method achieves SOTA performance in different orders of four tasks (A VE, A VVP, A VS and A VQA).

Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse domains, but developing high-performing models for specialized applications often requires substantial human annotation -- a process that is time-consuming, labor-intensive, and expensive. In this paper, we address the label-efficient learning problem for supervised finetuning (SFT) by leveraging task-diversity as a fundamental principle for effective data selection. This is markedly different from existing methods based on the prompt-diversity. Our approach is based on two key observations: 1) task labels for different prompts are often readily available; 2) pre-trained models have significantly varying levels of confidence across tasks. We combine these facts to devise a simple yet effective sampling strategy: we select examples across tasks using an inverse confidence weighting strategy. This produces models comparable to or better than those trained with more complex sampling procedures, while being significantly easier to implement and less computationally intensive. Notably, our experimental results demonstrate that this method can achieve better accuracy than training on the complete dataset (a 4\% increase in MMLU score). Across various annotation budgets and two instruction finetuning datasets, our algorithm consistently performs at or above the level of the best existing methods, while reducing annotation costs by up to 80\%.

Large language models (LLMs) are increasingly expected to function as collaborative partners, engaging in back-and-forth dialogue to solve complex, ambiguous problems. However, current LLMs often falter in real-world settings, defaulting to passive responses or narrow clarifications when faced with incomplete or under-specified prompts, falling short of proactively gathering the missing information that is crucial for high-quality solutions. In this work, we introduce a new task paradigm: proactive information gathering, where LLMs must identify gaps in the provided context and strategically elicit implicit user knowledge through targeted questions. To systematically study and train this capability, we design a scalable framework that generates partially specified, real-world tasks, masking key information and simulating authentic ambiguity. Within this setup, our core innovation is a reinforcement finetuning strategy that rewards questions that elicit genuinely new, implicit user information -- such as hidden domain expertise or fine-grained requirements -- that would otherwise remain unspoken. Experiments demonstrate that our trained Qwen-2.5-7B model significantly outperforms o3-mini by 18% on automatic evaluation metrics. More importantly, human evaluation reveals that clarification questions and final outlines generated by our model are favored by human annotators by 42% and 28% respectively. Together, these results highlight the value of proactive clarification in elevating LLMs from passive text generators to genuinely collaborative thought partners.

The distribution of data in the world (eg, internet, etc.) significantly differs from the well-curated datasets and is often over-populated with samples from common categories. The algorithms designed for well-curated datasets perform suboptimally when used for learning from imperfect datasets with long-tailed imbalances and distribution shifts. To expand the use of deep models, it is essential to overcome the labor-intensive curation process by developing robust algorithms that can learn from diverse, real-world data distributions. Toward this goal, we develop practical algorithms for Deep Neural Networks which can learn from limited and imperfect data present in the real world. This thesis is divided into four segments, each covering a scenario of learning from limited or imperfect data. The first part of the thesis focuses on Learning Generative Models from Long-Tail Data, where we mitigate the mode-collapse and enable diverse aesthetic image generations for tail (minority) classes. In the second part, we enable effective generalization on tail classes through Inductive Regularization schemes, which allow tail classes to generalize as effectively as the head classes without requiring explicit generation of images. In the third part, we develop algorithms for Optimizing Relevant Metrics for learning from long-tailed data with limited annotation (semi-supervised), followed by the fourth part, which focuses on the Efficient Domain Adaptation of the model to various domains with very few to zero labeled samples.

Progress in complex reasoning is constrained by the static nature of the current training datasets. We propose structured interaction as a new scaling axis, moving beyond the prevailing paradigm of increasing model parameters. Our self-evolving framework, AGORA, enables a collaborative ensemble to achieve reasoning performance exceeding state-of-the-art monolithic systems by up to 4.45 percentage points on challenging mathematical benchmarks. This gain stems from group emergent ability--the synthesis of collective capabilities unattainable by isolated models, validating interaction as a scalable driver of intelligence.

MOOCs offer free and open access to a wide audience, but completion rates remain low, often due to a lack of personalized content. To address this issue, it is essential to predict learner performance in order to provide tailored feedback. Behavioral traces-such as clicks and events-can be analyzed as time series to anticipate learners' outcomes. This work compares multivariate time series classification methods to identify at-risk learners at different stages of the course (after 5, 10 weeks, etc.). The experimental evaluation, conducted on the Open University Learning Analytics Dataset (OULAD), focuses on three courses: two in STEM and one in SHS. Preliminary results show that the evaluated approaches are promising for predicting learner failure in MOOCs. The analysis also suggests that prediction accuracy is influenced by the amount of recorded interactions, highlighting the importance of rich and diverse behavioral data.

A handful of college students who were part of OpenAI's testing cohort--hailing from Princeton, Wharton, and the University of Minnesota--shared positive reviews of Study Mode, saying it did a good job of checking their understanding and adapting to their pace. The learning approaches that OpenAI has programmed into Study Mode, which are based partially on Socratic methods, appear sound, says Christopher Harris, an educator in New York who has created a curriculum aimed at AI literacy. They might grant educators more confidence about allowing, or even encouraging, their students to use AI. "Professors will see this as working with them in support of learning as opposed to just being a way for students to cheat on assignments," he says. As demonstrated in OpenAI's recent partnership with leading teachers' unions, the company is currently trying to rebrand chatbots as tools for personalized learning rather than cheating.

An example of how'study mode' would work. Experts say it is'especially useful' for homework help, test prep and learning new topics It also features knowledge checks in the form of quizzes and open–ended questions, along with personalised feedback. The mode can also easy be toggled on and off during a conversation. Those wanting to use it should select'Study and learn' from tools in ChatGPT. 'Instead of doing the work for them, study mode encourages students to think critically about their learning', Robbie Torney, senior director of AI Programs at Common Sense Media said.