Each year, tens of millions of essays are written and graded in college-level English courses. Students are asked to analyze literary and cultural texts through a process known as close reading, in which they gather textual details to formulate evidence-based arguments. Despite being viewed as a basis for critical thinking and widely adopted as a required element of university coursework, close reading has never been evaluated on large language models (LLMs), and multi-discipline benchmarks like MMLU do not include literature as a subject. To fill this gap, we present KRISTEVA, the first close reading benchmark for evaluating interpretive reasoning, consisting of 1331 multiple-choice questions adapted from classroom data. With KRISTEVA, we propose three progressively more difficult sets of tasks to approximate different elements of the close reading process, which we use to test how well LLMs may seem to understand and reason about literary works: 1) extracting stylistic features, 2) retrieving relevant contextual information from parametric knowledge, and 3) multi-hop reasoning between style and external contexts. Our baseline results find that, while state-of-the-art LLMs possess some college-level close reading competency (accuracy 49.7% - 69.7%), their performances still trail those of experienced human evaluators on 10 out of our 11 tasks.

Good Job is Slate's advice column on work. Have a workplace problem big or small? I have been an eighth-grade teacher for seven years now and am beginning to think I made a terrible mistake in terms of choosing my profession. The kids I teach are rude and feral. They refuse to read or treat others with the slightest bit of decency, give up at the first sign of difficulty, and possess the attention span of goldfish.

Around midnight on April 16, 2025, after Chen Zimo learned that the Department of Homeland Security had threatened to revoke Harvard University's certification to enroll international students, he began communicating with a trusted source about possible legal scenarios. Chen, a Chinese citizen, still needed a number of courses before he could complete his degree in computer science at Harvard, and he felt panicked about the possibility of having his visa revoked. For him, the Harvard experience had been transformative. Chen--not his real name--had grown up in provincial China, where his family had modest resources and sent him to public schools. He could never have afforded Harvard without the university's generous financial support, and he had also received funding for summer language study.

Prompt-based text embedding models, which generate task-specific embeddings upon receiving tailored prompts, have recently demonstrated remarkable performance. However, their resulting embeddings often have thousands of dimensions, leading to high storage costs and increased computational costs of embedding-based operations. In this paper, we investigate how post-hoc dimensionality reduction applied to the embeddings affects the performance of various tasks that leverage these embeddings, specifically classification, clustering, retrieval, and semantic textual similarity (STS) tasks. Our experiments show that even a naive dimensionality reduction, which keeps only the first 25% of the dimensions of the embeddings, results in a very slight performance degradation, indicating that these embeddings are highly redundant. Notably, for classification and clustering, even when embeddings are reduced to less than 0.5% of the original dimensionality the performance degradation is very small. To quantitatively analyze this redundancy, we perform an analysis based on the intrinsic dimensionality and isotropy of the embeddings. Our analysis reveals that embeddings for classification and clustering, which are considered to have very high dimensional redundancy, exhibit lower intrinsic dimensionality and less isotropy compared with those for retrieval and STS.

We introduce Flashback Learning (FL), a novel method designed to harmonize the stability and plasticity of models in Continual Learning (CL). Unlike prior approaches that primarily focus on regularizing model updates to preserve old information while learning new concepts, FL explicitly balances this trade-off through a bidirectional form of regularization. This approach effectively guides the model to swiftly incorporate new knowledge while actively retaining its old knowledge. FL operates through a two-phase training process and can be seamlessly integrated into various CL methods, including replay, parameter regularization, distillation, and dynamic architecture techniques. In designing FL, we use two distinct knowledge bases: one to enhance plasticity and another to improve stability. FL ensures a more balanced model by utilizing both knowledge bases to regularize model updates. Theoretically, we analyze how the FL mechanism enhances the stability-plasticity balance. Empirically, FL demonstrates tangible improvements over baseline methods within the same training budget. By integrating FL into at least one representative baseline from each CL category, we observed an average accuracy improvement of up to 4.91% in Class-Incremental and 3.51% in Task-Incremental settings on standard image classification benchmarks. Additionally, measurements of the stability-to-plasticity ratio confirm that FL effectively enhances this balance. FL also outperforms state-of-the-art CL methods on more challenging datasets like ImageNet. Introduction Our brain excels at learning new information without significantly disrupting or forgetting previously acquired knowledge. In contrast, Deep Neural Networks (DNNs) often struggle with Catastrophic Forgetting (CF) [1], where new information can overwrite and interfere with existing knowledge. This phenomenon poses a significant challenge when continuous learning of new tasks is required. Continual Learning (CL) methods have been developed to mitigate catastrophic forgetting by enabling DNNs to retain prior knowledge while acquiring new skills. Deploying AI models in realistic settings demands that models function effectively when encountering data with varying distributions over time.

Educators teaching entry-level university engineering modules face the challenge of identifying which topics students find most difficult and how to support diverse student needs effectively. This study demonstrates a rigorous yet interpretable statistical approach -- hierarchical Bayesian modeling -- that leverages detailed student response data to quantify both skill difficulty and individual student abilities. Using a large-scale dataset from an undergraduate Statics course, we identified clear patterns of skill mastery and uncovered distinct student subgroups based on their learning trajectories. Our analysis reveals that certain concepts consistently present challenges, requiring targeted instructional support, while others are readily mastered and may benefit from enrichment activities. Importantly, the hierarchical Bayesian method provides educators with intuitive, reliable metrics without sacrificing predictive accuracy. This approach allows for data-informed decisions, enabling personalized teaching strategies to improve student engagement and success. By combining robust statistical methods with clear interpretability, this study equips educators with actionable insights to better support diverse learner populations.

Reinforcement Learning (RL) has demonstrated tremendous empirical success across numerous challenging domains. However, we lack a strong theoretical understanding of the statistical complexity of RL in environments with large state spaces, where function approximation is required for sample-efficient learning. This thesis addresses this gap by rigorously examining the statistical complexity of RL with function approximation from a learning theoretic perspective. Departing from a long history of prior work, we consider the weakest form of function approximation, called agnostic policy learning, in which the learner seeks to find the best policy in a given class $Π$, with no guarantee that $Π$ contains an optimal policy for the underlying task. We systematically explore agnostic policy learning along three key axes: environment access -- how a learner collects data from the environment; coverage conditions -- intrinsic properties of the underlying MDP measuring the expansiveness of state-occupancy measures for policies in the class $Π$, and representational conditions -- structural assumptions on the class $Π$ itself. Within this comprehensive framework, we (1) design new learning algorithms with theoretical guarantees and (2) characterize fundamental performance bounds of any algorithm. Our results reveal significant statistical separations that highlight the power and limitations of agnostic policy learning.

In orthogonal frequency division multiplexing (OFDM), accurate channel estimation is crucial. Classical signal processing based approaches, such as minimum mean-squared error (MMSE) estimation, often require second-order statistics that are difficult to obtain in practice. Recent deep neural networks based methods have been introduced to address this; yet they often suffer from high complexity. This paper proposes an Attention-aided MMSE (A-MMSE), a novel model-based DNN framework that learns the optimal MMSE filter via the Attention Transformer. Once trained, the A-MMSE estimates the channel through a single linear operation for channel estimation, eliminating nonlinear activations during inference and thus reducing computational complexity. To enhance the learning efficiency of the A-MMSE, we develop a two-stage Attention encoder, designed to effectively capture the channel correlation structure. Additionally, a rank-adaptive extension of the proposed A-MMSE allows flexible trade-offs between complexity and channel estimation accuracy. Extensive simulations with 3GPP TDL channel models demonstrate that the proposed A-MMSE consistently outperforms other baseline methods in terms of normalized MSE across a wide range of SNR conditions. In particular, the A-MMSE and its rank-adaptive extension establish a new frontier in the performance complexity trade-off, redefining the standard for practical channel estimation methods.

We study online and transductive online learning when the learner interacts with the concept class only via Empirical Risk Minimization (ERM) or weak consistency oracles on arbitrary instance subsets. This contrasts with standard online models, where the learner knows the entire class. The ERM oracle returns a hypothesis minimizing loss on a given subset, while the weak consistency oracle returns a binary signal indicating whether the subset is realizable by some concept. The learner is evaluated by the number of mistakes and oracle calls. In the standard online setting with ERM access, we prove tight lower bounds in both realizable and agnostic cases: $Ω(2^{d_{VC}})$ mistakes and $Ω(\sqrt{T 2^{d_{LD}}})$ regret, where $T$ is the number of timesteps and $d_{LD}$ is the Littlestone dimension. We further show that existing online learning results with ERM access carry over to the weak consistency setting, incurring an additional $O(T)$ in oracle calls. We then consider the transductive online model, where the instance sequence is known but labels are revealed sequentially. For general Littlestone classes, we show that optimal realizable and agnostic mistake bounds can be achieved using $O(T^{d_{VC}+1})$ weak consistency oracle calls. On the negative side, we show that limiting the learner to $Ω(T)$ weak consistency queries is necessary for transductive online learnability, and that restricting the learner to $Ω(T)$ ERM queries is necessary to avoid exponential dependence on the Littlestone dimension. Finally, for certain concept classes, we reduce oracle calls via randomized algorithms while maintaining similar mistake bounds. In particular, for Thresholds on an unknown ordering, $O(\log T)$ ERM queries suffice; for $k$-Intervals, $O(T^3 2^{2k})$ weak consistency queries suffice.

This paper introduces DLM-One, a score-distillation-based framework for one-step sequence generation with continuous diffusion language models (DLMs). DLM-One eliminates the need for iterative refinement by aligning the scores of a student model's outputs in the continuous token embedding space with the score function of a pretrained teacher DLM. We investigate whether DLM-One can achieve substantial gains in sampling efficiency for language modeling. Through comprehensive experiments on DiffuSeq -- a representative continuous DLM -- we show that DLM-One achieves up to ~500x speedup in inference time while maintaining competitive performance on benchmark text generation tasks used to evaluate the teacher models. We further analyze the method's empirical behavior across multiple datasets, providing initial insights into its generality and practical applicability. Our findings position one-step diffusion as a promising direction for efficient, high-quality language generation and broader adoption of continuous diffusion models operating in embedding space for natural language processing.