For reliable large-scale quantum computation, a quantum error correction (QEC) scheme must effectively resolve physical errors to protect logical information. Leveraging recent advances in deep learning, neural network-based decoders have emerged as a promising approach to enhance the reliability of QEC. We propose the Hierarchical Qubit-Merging Transformer (HQMT), a novel and general decoding framework that explicitly leverages the structural graph of stabilizer codes to learn error correlations across multiple scales. Our architecture first computes attention locally on structurally related groups of stabilizers and then systematically merges these qubit-centric representations to build a global view of the error syndrome. The proposed HQMT achieves substantially lower logical error rates for surface codes by integrating a dedicated qubit-merging layer within the transformer architecture. Across various code distances, HQMT significantly outperforms previous neural network-based QEC decoders as well as a powerful belief propagation with ordered statistics decoding (BP+OSD) baseline. This hierarchical approach provides a scalable and effective framework for surface code decoding, advancing the realization of reliable quantum computing.

Many in-silico simulations of human survey responses with large language models (LLMs) focus on generating closed-ended survey responses, whereas LLMs are typically trained to generate open-ended text instead. Previous research has used a diverse range of methods for generating closed-ended survey responses with LLMs, and a standard practice remains to be identified. In this paper, we systematically investigate the impact that various Survey Response Generation Methods have on predicted survey responses. We present the results of 32 mio. simulated survey responses across 8 Survey Response Generation Methods, 4 political attitude surveys, and 10 open-weight language models. We find significant differences between the Survey Response Generation Methods in both individual-level and subpopulation-level alignment. Our results show that Restricted Generation Methods perform best overall, and that reasoning output does not consistently improve alignment. Our work underlines the significant impact that Survey Response Generation Methods have on simulated survey responses, and we develop practical recommendations on the application of Survey Response Generation Methods.

Existing benchmarks that measure cultural adaptation in LLMs are misaligned with the actual challenges these models face when interacting with users from diverse cultural backgrounds. In this work, we introduce the first framework and benchmark designed to evaluate LLMs in realistic, multicultural conversational settings. Grounded in sociocultural theory, our framework formalizes how linguistic style - a key element of cultural communication - is shaped by situational, relational, and cultural context. We construct a benchmark dataset based on this framework, annotated by culturally diverse raters, and propose a new set of desiderata for cross-cultural evaluation in NLP: conversational framing, stylistic sensitivity, and subjective correctness. We evaluate today's top LLMs on our benchmark and show that these models struggle with cultural adaptation in a conversational setting.

The advent of LLMs has given rise to a new type of web search: Generative search, where LLMs retrieve web pages related to a query and generate a single, coherent text as a response. This output modality stands in stark contrast to traditional web search, where results are returned as a ranked list of independent web pages. In this paper, we ask: Along what dimensions do generative search outputs differ from traditional web search? We compare Google, a traditional web search engine, with four generative search engines from two providers (Google and OpenAI) across queries from four domains. Our analysis reveals intriguing differences. Most generative search engines cover a wider range of sources compared to web search. Generative search engines vary in the degree to which they rely on internal knowledge contained within the model parameters v.s. external knowledge retrieved from the web. Generative search engines surface varying sets of concepts, creating new opportunities for enhancing search diversity and serendipity. Our results also highlight the need for revisiting evaluation criteria for web search in the age of Generative AI.

Governance of data, compliance, and business privacy matters, particularly for healthcare and finance businesses. Since the recent emergence of AI enterprise AI assistants enhancing business productivity, safeguarding private data and compliance is now a priority. With the implementation of AI assistants across the enterprise, the zero data retention can be achieved by implementing zero data retention policies by Large Language Model businesses like Open AI and Anthropic and Meta. In this work, we explore zero data retention policies for the Enterprise apps of large language models (LLMs). Our key contribution is defining the architectural, compliance, and usability trade-offs of such systems in parallel. In this research work, we examine the development of commercial AI assistants with two industry leaders and market titans in this arena - Salesforce and Microsoft. Both of these companies used distinct technical architecture to support zero data retention policies. Salesforce AgentForce and Microsoft Copilot are among the leading AI assistants providing much-needed push to business productivity in customer care. The purpose of this paper is to analyze the technical architecture and deployment of zero data retention policy by consuming applications as well as big language models service providers like Open Ai, Anthropic, and Meta.

Retrieval-augmented generation (RAG) has demonstrated its ability to enhance Large Language Models (LLMs) by integrating external knowledge sources. However, multi-hop questions, which require the identification of multiple knowledge targets to form a synthesized answer, raise new challenges for RAG systems. Under the multi-hop settings, existing methods often struggle to fully understand the questions with complex semantic structures and are susceptible to irrelevant noise during the retrieval of multiple information targets. To address these limitations, we propose a novel graph representation learning framework for multi-hop question retrieval. We first introduce a Multi-information Level Knowledge Graph (Multi-L KG) to model various information levels for a more comprehensive understanding of multi-hop questions. Based on this, we design a Query-Specific Graph Neural Network (QSGNN) for representation learning on the Multi-L KG. QSGNN employs intra/inter-level message passing mechanisms, and in each message passing the information aggregation is guided by the query, which not only facilitates multi-granular information aggregation but also significantly reduces the impact of noise. To enhance its ability to learn robust representations, we further propose two synthesized data generation strategies for pre-training the QSGNN. Extensive experimental results demonstrate the effectiveness of our framework in multi-hop scenarios, especially in high-hop questions the improvement can reach 33.8\%. The code is available at: https://github.com/Jerry2398/QSGNN.

We propose a novel neural architecture named TextGraphFuseGAT, which integrates a pretrained transformer encoder (PhoBERT) with Graph Attention Networks for token-level classification tasks. The proposed model constructs a fully connected graph over the token embeddings produced by PhoBERT, enabling the GAT layer to capture rich inter-token dependencies beyond those modeled by sequential context alone. To further enhance contextualization, a Transformer-style self-attention layer is applied on top of the graph-enhanced embeddings. The final token representations are passed through a classification head to perform sequence labeling. We evaluate our approach on three Vietnamese benchmark datasets: PhoNER-COVID19 for named entity recognition in the COVID-19 domain, PhoDisfluency for speech disfluency detection, and VietMed-NER for medical-domain NER. VietMed-NER is the first Vietnamese medical spoken NER dataset, featuring 18 entity types collected from real-world medical speech transcripts and annotated with the BIO tagging scheme. Its specialized vocabulary and domain-specific expressions make it a challenging benchmark for token-level classification models. Experimental results show that our method consistently outperforms strong baselines, including transformer-only and hybrid neural models such as BiLSTM + CNN + CRF, confirming the effectiveness of combining pretrained semantic features with graph-based relational modeling for improved token classification across multiple domains.

Research on reasoning in language models (LMs) predominantly focuses on improving the correctness of their outputs. But some important applications require modeling reasoning patterns that are incorrect. For example, automated systems that can reason about and simulate student errors are useful for providing real-time feedback in the classroom or offline practice for educators-in-training. This paper presents a new method, MISTAKE, that (1) constructs high-quality synthetic examples of reasoning errors by leveraging cycle consistency between incorrect answers and latent misconceptions; and (2) uses the generated data to learn models for student simulation, misconception classification, and answer generation. We evaluate MISTAKE on three educational tasks and find that it results in (1) higher accuracy when simulating incorrect student answers based on specific misconceptions, (2) increased performance inferring latent misconceptions from observed incorrect answers, and (3) higher alignment with expert-written distractor answers when generating incorrect answers (e.g., for multiple-choice tests).

While Large Language Models (LLMs) excel at algorithmic code generation, they struggle with front-end development, where correctness is judged on rendered pixels and interaction. We present ReLook, an agentic, vision-grounded reinforcement learning framework that empowers an agent to close a robust generate--diagnose--refine loop by invoking a multimodal LLM (MLLM) as a tool. During training, the agent uses the MLLM-in-the-loop both as a visual critic--scoring code with screenshots--and as a source of actionable, vision-grounded feedback; a strict zero-reward rule for invalid renders anchors renderability and prevents reward hacking. To prevent behavioral collapse, we introduce Forced Optimization, a strict acceptance rule that admits only improving revisions, yielding monotonically better trajectories. At inference, we decouple the critic and run a lightweight, critic-free self-edit cycle, keeping latency comparable to base decoding while retaining most of the gains. Across three widely used benchmarks, ReLook consistently outperforms strong baselines in vision-grounded front-end code generation, highlighting the benefits of agentic perception, visual rewards, and training-inference decoupling.

Integer AI inference significantly reduces computational complexity in embedded systems. Quantization-aware training (QAT) helps mitigate accuracy degradation associated with post-training quantization but still overlooks the impact of integer rescaling during inference, which is a hardware costly operation in integer-only AI inference. This work shows that rescaling cost can be dramatically reduced post-training, by applying a stronger quantization to the rescale multiplicands at no model-quality loss. Furthermore, we introduce Rescale-Aware Training, a fine tuning method for ultra-low bit-width rescaling multiplicands. Experiments show that even with 8x reduced rescaler widths, the full accuracy is preserved through minimal incremental retraining. This enables more energy-efficient and cost-efficient AI inference for resource-constrained embedded systems.