Successful communication depends on the speaker's intended style (i.e., what the speaker is trying to convey) aligning with the listener's interpreted style (i.e., what the listener perceives).However, cultural differences often lead to misalignment between the two; for example, politeness is often lost in translation.We characterize the ways that LLMs fail to translate style -biasing translations towards neutrality and performing worse in non-Western languages.We mitigate these failures with RASTA (Retrieval-Augmented STylistic Alignment), a method that leverages learned stylistic concepts to encourage LLM translation to appropriately convey cultural communication norms and align style. 1 * Equal contribution 1 Code, data, & prompts available at https://github. com/shreyahavaldar/style_alignment "After class today, I chatted with Eric, my professor." 今日の授 業の後、私は教授のエリックとお しゃべりしました。 Chat

Large language models (LLMs) excel at zero-shot inference but continue to struggle with complex, multi-step reasoning. Recent methods that augment LLMs with intermediate reasoning steps such as Chain of Thought (CoT) and Program of Thought (PoT) improve performance but often produce undesirable solutions, especially in algorithmic domains. We introduce Per-Instance Program Synthesis (PIPS), a method that generates and refines programs at the instance level using structural feedback without relying on task-specific guidance or explicit test cases. To further improve performance, PIPS incorporates a confidence metric that dynamically chooses between direct inference and program synthesis on a per-instance basis. Experiments across three frontier LLMs and 30 benchmarks including all tasks of Big Bench Extra Hard (BBEH), visual question answering tasks, relational reasoning tasks, and mathematical reasoning tasks show that PIPS improves the absolute harmonic mean accuracy by up to 8.6\% and 9.4\% compared to PoT and CoT respectively, and reduces undesirable program generations by 65.1\% on the algorithmic tasks compared to PoT with Gemini-2.0-Flash.

Concept-based interpretability methods offer a lens into the internals of foundation models by decomposing their embeddings into high-level concepts. These concept representations are most useful when they are *compositional*, meaning that the individual concepts compose to explain the full sample. We show that existing unsupervised concept extraction methods find concepts which are not compositional. To automatically discover compositional concept representations, we identify two salient properties of such representations, and propose Compositional Concept Extraction (CCE) for finding concepts which obey these properties. We evaluate CCE on five different datasets over image and text data. Our evaluation shows that CCE finds more compositional concept representations than baselines and yields better accuracy on four downstream classification tasks.

Finding errors in machine learning applications requires a thorough exploration of their behavior over data. Existing approaches used by practitioners are often ad-hoc and lack the abstractions needed to scale this process. We present TorchQL, a programming framework to evaluate and improve the correctness of machine learning applications. TorchQL allows users to write queries to specify and check integrity constraints over machine learning models and datasets. It seamlessly integrates relational algebra with functional programming to allow for highly expressive queries using only eight intuitive operators. We evaluate TorchQL on diverse use-cases including finding critical temporal inconsistencies in objects detected across video frames in autonomous driving, finding data imputation errors in time-series medical records, finding data labeling errors in real-world images, and evaluating biases and constraining outputs of language models. Our experiments show that TorchQL enables up to 13x faster query executions than baselines like Pandas and MongoDB, and up to 40% shorter queries than native Python. We also conduct a user study and find that TorchQL is natural enough for developers familiar with Python to specify complex integrity constraints.

Sample re-weighting strategies provide a promising mechanism to deal with imperfect training data in machine learning, such as noisily labeled or class-imbalanced data. One such strategy involves formulating a bi-level optimization problem called the meta re-weighting problem, whose goal is to optimize performance on a small set of perfect pivotal samples, called meta samples. Many approaches have been proposed to efficiently solve this problem. However, all of them assume that a perfect meta sample set is already provided while we observe that the selections of meta sample set is performance-critical. In this paper, we study how to learn to identify such a meta sample set from a large, imperfect training set, that is subsequently cleaned and used to optimize performance in the meta re-weighting setting. We propose a learning framework which reduces the meta samples selection problem to a weighted K-means clustering problem through rigorously theoretical analysis. We propose two clustering methods within our learning framework, Representation-based clustering method (RBC) and Gradient-based clustering method (GBC), for balancing performance and computational efficiency. Empirical studies demonstrate the performance advantage of our methods over various baseline methods

Qing Lyu, Shreya Havaldar, Adam Stein, Li Zhang, Delip Rao, Eric Wong, Marianna Apidianaki, Chris Callison-Burch. Proceedings of the 13th International Joint Conference on Natural Language Processing and the 3rd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

We study the problem of synthesizing a core fragment of relational queries called select-project-join (SPJ) queries from input-output examples. Search-based synthesis techniques are suited to synthesizing projections and joins by navigating the network of relational tables but require additional supervision for synthesizing comparison predicates. On the other hand, decision tree learning techniques are suited to synthesizing comparison predicates when the input database can be summarized as a single labelled relational table. In this paper, we adapt and interleave methods from the domains of relational query synthesis and decision tree learning, and present an end-to-end framework for synthesizing relational queries with categorical and numerical comparison predicates. Our technique guarantees the completeness of the synthesis procedure and strongly encourages minimality of the synthesized program. We present Libra, an implementation of this technique and evaluate it on a benchmark suite of 1,475 instances of queries over 159 databases with multiple tables. Libra solves 1,361 of these instances in an average of 59 seconds per instance. It outperforms state-of-the-art program synthesis tools Scythe and PatSQL in terms of both the running time and the quality of the synthesized programs.

Measuring the causal effect of advertising on driving desired behavior is an important problem to the digital publishing industry (the "attribution" problem). It is common to use observational methods for attribution, due to the high cost and difficulty of employing randomized controlled trials (RCTs). However, recent results have shown that even current sophisticated observational methods may be inaccurate, yielding incorrect estimates of the true effect of advertising. Here, we present a new observational attribution method based on a successful model of neural spiking that learns the temporal interactions between event-based time series. We train this model on data from several RCT marketing experiments, and show that it can accurately recover the true causal attribution.