While data-driven methods such as neural operator have achieved great success in solving differential equations (DEs), they suffer from domain shift problems caused by different learning environments (with data bias or equation changes), which can be alleviated by transfer learning (TL). However, existing TL methods adopted in DEs problems lack either generalizability in general DEs problems or physics preservation during training. In this work, we focus on a general transfer learning method that adaptively correct the domain shift and preserve physical information. Mathematically, we characterize the data domain as product distribution and the essential problems as distribution bias and operator bias. A Physics-preserved Optimal Tensor Transport (POTT) method that simultaneously admits generalizability to common DEs and physics preservation of specific problem is proposed to adapt the data-driven model to target domain utilizing the push-forward distribution induced by the POTT map. Extensive experiments demonstrate the superior performance, generalizability and physics preservation of the proposed POTT method.

Training large language models (LLMs) as interactive agents presents unique challenges including long-horizon decision making and interacting with stochastic environment feedback. While reinforcement learning (RL) has enabled progress in static tasks, multi-turn agent RL training remains underexplored. We propose StarPO (State-Thinking-Actions-Reward Policy Optimization), a general framework for trajectory-level agent RL, and introduce RAGEN, a modular system for training and evaluating LLM agents. Our study on four stylized environments reveals three core findings. First, our agent RL training shows a recurring mode of Echo Trap where reward variance cliffs and gradient spikes; we address this with StarPO-S, a stabilized variant with trajectory filtering, critic incorporation, and gradient stabilization. Second, we find the shaping of RL rollouts would benefit from diverse initial states, medium interaction granularity and more frequent sampling. Third, we show that without fine-grained, reasoning-aware reward signals, agent reasoning hardly emerge through multi-turn RL and they may show shallow strategies or hallucinated thoughts. Code and environments are available at https://github.com/RAGEN-AI/RAGEN.

Knowledge distillation is a widely used technique for compressing large language models (LLMs), in which a smaller student model is trained to mimic a larger teacher model. Typically, both the teacher and student models are Transformer-based architectures, leveraging softmax attention for sequence modeling. However, the quadratic complexity of self-attention during inference remains a significant bottleneck, motivating the exploration of subquadratic alternatives such as structured state-space models (SSMs), linear attention, and recurrent architectures. In this work, we systematically evaluate the transferability of knowledge distillation from a Transformer teacher model to eight subquadratic student architectures. Our study investigates which subquadratic model can most effectively approximate the teacher model's learned representations through knowledge distillation, and how different architectural design choices influence the training dynamics. We further investigate the impact of initialization strategies, such as matrix mixing and query-key-value (QKV) copying, on the adaptation process. Our empirical results on multiple NLP benchmarks provide insights into the trade-offs between efficiency and performance, highlighting key factors for successful knowledge transfer to subquadratic architectures.

Semi-Supervised Semantic Segmentation (SSSS) aims to improve segmentation accuracy by leveraging a small set of labeled images alongside a larger pool of unlabeled data. Recent advances primarily focus on pseudo-labeling, consistency regularization, and co-training strategies. However, existing methods struggle to balance global semantic representation with fine-grained local feature extraction. To address this challenge, we propose a novel tri-branch semi-supervised segmentation framework incorporating a dual-teacher strategy, named IGL-DT. Our approach employs SwinUnet for high-level semantic guidance through Global Context Learning and ResUnet for detailed feature refinement via Local Regional Learning. Additionally, a Discrepancy Learning mechanism mitigates over-reliance on a single teacher, promoting adaptive feature learning. Extensive experiments on benchmark datasets demonstrate that our method outperforms state-of-the-art approaches, achieving superior segmentation performance across various data regimes.

The ability of a model to learn a task depends strongly on both the task difficulty and the model size. We aim to understand how task difficulty relates to the minimum number of parameters required for learning specific tasks in small transformer models. Our study focuses on the ListOps dataset, which consists of nested mathematical operations. We gradually increase task difficulty by introducing new operations or combinations of operations into the training data. We observe that sum modulo n is the hardest to learn. Curiously, when combined with other operations such as maximum and median, the sum operation becomes easier to learn and requires fewer parameters. We show that joint training not only improves performance but also leads to qualitatively different model behavior. We show evidence that models trained only on SUM might be memorizing and fail to capture the number structure in the embeddings. In contrast, models trained on a mixture of SUM and other operations exhibit number-like representations in the embedding space, and a strong ability to distinguish parity. Furthermore, the SUM-only model relies more heavily on its feedforward layers, while the jointly trained model activates the attention mechanism more. Finally, we show that learning pure SUM can be induced in models below the learning threshold of pure SUM, by pretraining them on MAX+MED. Our findings indicate that emergent abilities in language models depend not only on model size, but also the training curriculum.

Despite advances in designing personas for Large Language Models (LLM), challenges remain in aligning them with human cognitive processes and representing diverse stakeholder perspectives. We introduce a Social Cognitive Theory (SCT) agent design framework for designing, evaluating, and implementing psychologically grounded LLMs with consistent behavior. Our framework operationalizes SCT through four personal factors (cognitive, motivational, biological, and affective) for designing, six quantifiable constructs for evaluating, and a graph database-backed architecture for implementing stakeholder personas. Experiments tested agents' responses to contradicting information of varying reliability. In the highly polarized renewable energy transition discourse, we design five diverse agents with distinct ideologies, roles, and stakes to examine stakeholder representation. The evaluation of these agents in contradictory scenarios occurs through comprehensive processes that implement the SCT. Results show consistent response patterns ($R^2$ range: $0.58-0.61$) and systematic temporal development of SCT construct effects. Principal component analysis identifies two dimensions explaining $73$% of variance, validating the theoretical structure. Our framework offers improved explainability and reproducibility compared to black-box approaches. This work contributes to ongoing efforts to improve diverse stakeholder representation while maintaining psychological consistency in LLM personas.

Continual reinforcement learning (RL) concerns agents that are expected to learn continually, rather than converge to a policy that is then fixed for evaluation. Such an approach is well suited to environments the agent perceives as changing, which renders any static policy ineffective over time. The few simulators explicitly designed for empirical research in continual RL are often limited in scope or complexity, and it is now common for researchers to modify episodic RL environments by artificially incorporating abrupt task changes during interaction. In this paper, we introduce AgarCL, a research platform for continual RL that allows for a progression of increasingly sophisticated behaviour. AgarCL is based on the game Agar.io, a non-episodic, high-dimensional problem featuring stochastic, ever-evolving dynamics, continuous actions, and partial observability. Additionally, we provide benchmark results reporting the performance of DQN, PPO, and SAC in both the primary, challenging continual RL problem, and across a suite of smaller tasks within AgarCL, each of which isolates aspects of the full environment and allow us to characterize the challenges posed by different aspects of the game.

LLMs have been demonstrated to align with the values of Western or North American cultures. Prior work predominantly showed this effect through leveraging surveys that directly ask (originally people and now also LLMs) about their values. However, it is hard to believe that LLMs would consistently apply those values in real-world scenarios. To address that, we take a bottom-up approach, asking LLMs to reason about cultural norms in narratives from different cultures. We find that GPT-4 tends to generate norms that, while not necessarily incorrect, are significantly less culture-specific. In addition, while it avoids overtly generating stereotypes, the stereotypical representations of certain cultures are merely hidden rather than suppressed in the model, and such stereotypes can be easily recovered. Addressing these challenges is a crucial step towards developing LLMs that fairly serve their diverse user base.

The recent rise of Large Reasoning Models (LRMs) has significantly improved multi-step reasoning performance, but often at the cost of generating excessively long reasoning chains. This paper revisits the efficiency of such reasoning processes through an information-theoretic lens, revealing a fundamental trade-off between reasoning length and semantic efficiency. We propose two metrics, InfoBias and InfoGain, to quantify divergence from ideal reasoning paths and stepwise information contribution, respectively. Empirical analyses show that longer reasoning chains tend to exhibit higher information bias and diminishing information gain, especially for incorrect answers. Motivated by these findings, we introduce an entropy-based Adaptive Think strategy that dynamically halts reasoning once confidence is sufficiently high, improving efficiency while maintaining competitive accuracy. Compared to the Vanilla Think approach (default mode), our strategy yields a 1.10% improvement in average accuracy and a 50.80% reduction in token usage on QwQ-32B across six benchmark tasks spanning diverse reasoning types and difficulty levels, demonstrating superior efficiency and reasoning performance. These results underscore the promise of entropy-based methods for enhancing both accuracy and cost-effiiciency in large language model deployment.

The rapid advancement of Large Language Models (LLMs) has opened new avenues in education. This study examines the use of LLMs in supporting learning in machine learning education; in particular, it focuses on the ability of LLMs to identify common errors of practice (pitfalls) in machine learning code, and their ability to provide feedback that can guide learning. Using a portfolio of code samples, we consider four different LLMs: one closed model and three open models. Whilst the most basic pitfalls are readily identified by all models, many common pitfalls are not. They particularly struggle to identify pitfalls in the early stages of the ML pipeline, especially those which can lead to information leaks, a major source of failure within applied ML projects. They also exhibit limited success at identifying pitfalls around model selection, which is a concept that students often struggle with when first transitioning from theory to practice. This questions the use of current LLMs to support machine learning education, and also raises important questions about their use by novice practitioners. Nevertheless, when LLMs successfully identify pitfalls in code, they do provide feedback that includes advice on how to proceed, emphasising their potential role in guiding learners. We also compare the capability of closed and open LLM models, and find that the gap is relatively small given the large difference in model sizes. This presents an opportunity to deploy, and potentially customise, smaller more efficient LLM models within education, avoiding risks around cost and data sharing associated with commercial models.