arXiv subjects

Xinze Lyu

Publications and source records attributed to Xinze Lyu.

14 recordsLinked to original sources

arXiv2026

Retrieval-Enhanced Cultural Evaluation for Social Media Translation Training

Social media translation requires preserving meaning while capturing the cultural nuances of user-generated content (UGC). However, gaps in the cultural knowledge of LLM-based evaluators can lead to inaccurate assessments and unreliable training rewards. To improve these signals, we use retrieval-enhanced cultural evaluation to train CUE-MT, a family of social media translation models. A calibrated, domain-specific Judger guides supervised data refinement and is augmented by an Agent for hard-example mining and reinforcement learning. The Agent retrieves relevant external evidence as needed and distills it into concise hints that the Judger uses to score translations and assign rewards. Across 9B, 27B, and 35B-A3B backbones, CUE-MT outperforms the evaluated open-source baselines of comparable size on CUE-MT-Bench and three public benchmarks. Human evaluation on CUE-MT-Bench shows cultural effectiveness gains of 7.44-15.12 percentage points over the strongest external baseline in each size group. Retrieval improves the Judger's four-way classification accuracy by 8.07 percentage points over a matched no-retrieval baseline that uses LLM-generated hints alone. Training ablations further support the combination of hard-example mining and Agent-enhanced rewards.

cs.CL↗

arXiv2026

CULTURE-CMT: Benchmarking Contextualized Social Media Translation

Social-media comments are interactions, yet they are often translated and evaluated as isolated text. This mismatch is especially problematic when meaning depends on surrounding discussion, community-specific slang, or culturally situated expressions. We introduce CULTURE-CMT, a Chinese-English benchmark for contextualized social-media comment translation that restores the post and conversational context naturally available to users. Its 572 manually verified examples are organized into Slang, Context, and Both, enabling diagnostic analysis of distinct and interacting sources of translation difficulty. To support scalable evaluation, we further develop a retrieval-augmented cultural-effectiveness Judger that combines a calibrated scoring model with dynamically retrieved cultural knowledge. CULTURE-CMT reveals several non-trivial patterns. Providing interaction context improves most model configurations, but does not eliminate culturally ineffective translations. Examples requiring both contextual grounding and culturally situated language remain substantially more difficult than either source of difficulty alone. Moreover, larger models are not consistently better, and thinking produces highly model-dependent gains, including occasional degradation. In contrast, a 27B model adapted on benchmark-disjoint social-media data achieves the lowest observed invalidity rate among the evaluated configurations under thinking inference with context. Together, these results suggest that realistic social-media translation is not simply a matter of scaling models or reasoning longer, but of learning to translate meaning as it is constructed within evolving online interactions.

cs.CL↗

arXiv2026

Beyond Words Alone: Disentangling Distributed Evidence for Visually Grounded Cultural Translation

Social media enables cross-lingual communication, yet the meaning of user-generated content (UGC) is often distributed across text, images, and cultural knowledge, making translation challenging. Existing evaluations typically treat images as optional and assume textual information is sufficient, obscuring whether errors stem from missing visual evidence or inadequate cultural adaptation. We introduce \textbf{Culture-MMT}, a diagnostic benchmark for \textbf{distributed-evidence pragmatic reconstruction} in social-media translation. It contains 498 curated, de-identified UGC instances across 20 domains, categorized into Symbol, Visual, and Hybrid by their required evidence. We also develop a human-aligned domain-specific judge, Judger, and a controlled evaluation framework that systematically varies the evidence available to translators. Across 17 model configurations, real images improve performance by +0.280 on Visual and +0.197 on Hybrid, but only +0.050 on Symbol, while random images provide no comparable gains. Human-authored descriptions further show that jointly dependent cases remain more challenging even when visual facts are explicit. These results characterize culturally effective social-media translation as {meaning reconstruction from distributed evidence}, highlighting visual grounding and cultural-pragmatic adaptation as distinct model bottlenecks.

cs.CL↗

arXiv2026

Benchmarking Machine Translation on Chinese Social Media Texts

The prevalence of rapidly evolving slang, neologisms, and highly stylized expressions in informal user-generated text, particularly on Chinese social media, poses significant challenges for Machine Translation (MT) benchmarking. Specifically, we identify two primary obstacles: (1) data scarcity, as high-quality parallel data requires bilingual annotators familiar with platform-specific slang, and stylistic cues in both languages; and (2) metric limitations, where traditional evaluators like COMET often fail to capture stylistic fidelity and nonstandard expressions. To bridge these gaps, we introduce CSM-MTBench, a benchmark covering five Chinese-foreign language directions and consisting of two expert-curated subsets: Fun Posts, featuring context-rich, slang- and neologism-heavy content, and Social Snippets, emphasizing concise, emotion- and style- driven expressions. Furthermore, we propose tailored evaluation approaches for each subset: measuring the translation success rate of slang and neologisms in Fun Posts, while assessing tone and style preservation in Social Snippets via a hybrid of embedding-based metrics and LLM-as-a-judge. Experiments on over 20 models reveal substantial variation in how current MT systems handle semantic fidelity and informal, social-media-specific stylistic cues. CSM-MTBench thus serves as a rigorous testbed for advancing MT systems capable of mastering real-world Chinese social media texts.

cs.CL↗

arXiv2026

Beyond Literal Translation: Evaluating Cultural Effectiveness in Social Media UGC

Social media platforms enable large-scale cross-lingual communication, but translating user-generated content (UGC) remains challenging due to its informal style, cultural references, and interaction-based expressions. While recent LLMs have improved translation quality, existing benchmarks and metrics often fail to capture whether translations convey intended meaning and cultural resonance in real-world settings. In this work, we introduce CULTURE-MT, a benchmark for social media translation that focuses on both CULtural Transmission and UGC-specific emotion REsonance. CULTURE-MT consists of 1,002 UGC notes across 14 domains, categorized into four types based on culture-loaded symbols and linguistic style features. We also construct UGC-oriented training data to fine-tune Qwen3-8B and Qwen3-32B as baselines. We propose cultural effectiveness as a new evaluation criterion, focusing on expression accuracy and cultural adaptability. Testing 15 models, including the baselines, we find that traditional metrics fail to capture cultural effectiveness. We also observe that cultural effectiveness on base LLMs correlates with model size. Our work provides a comprehensive evaluation system for UGC translation models and will offer an open evaluation platform to advance research in this area. We release the CULTURE-MT benchmark and provide an online leaderboard where submitted translation results can be evaluated by our trained JUDGER.

cs.CL↗

arXiv2026

Rate-Splitting Multiple Access with a SIC-Free Receiver: An Experimental Study

Most Rate-Splitting Multiple Access (RSMA) implementations rely on successive interference cancellation (SIC) at the receiver, whose performance is inherently limited by error propagation during common-stream decoding. This paper addresses this issue by developing a SIC-free RSMA receiver based on joint demapping (JD), which directly evaluates bit vectors over a composite constellation. Using a two-user Multiple-Input Single-Output (MISO) prototype, we conduct over-the-air measurements to systematically compare SIC and JD-based receivers. The results show that the proposed SIC-free receiver provides stronger reliability and better practicality over a wider operating range, with all observations being consistent with theoretical expectations.

eess.SP↗

arXiv2025

Rate-Splitting Multiple Access for Integrated Sensing and Communications: A First Experimental Study

A canonical use case of Integrated Sensing and Communications (ISAC) in multiple-input multiple-output (MIMO) systems involves a multi-antenna transmitter communicating with $K$ users and sensing targets in its vicinity. For this setup, precoder and multiple access designs are of utmost importance, as the limited transmit power budget must be efficiently directed towards the desired directions (users and targets) to maximize both communications and sensing performance. This problem has been widely investigated analytically under various design choices, in particular (a) whether or not a dedicated sensing signal is needed, and (b) for different MIMO multiple access techniques, such as Space Division Multiple Access (SDMA) and Rate-Splitting Multiple Access (RSMA). However, a conclusive answer on which design choice achieves the best ISAC performance, backed by experimental results, remains elusive. We address this vacuum by experimentally evaluating and comparing RSMA and SDMA for communicating with two users $(K = 2)$ and sensing (ranging) one target. Over three scenarios that are representative of \emph{vehicular} ISAC, covering different levels of inter-user interference and separation/integration between sensing and communications, we show that RSMA without a dedicated sensing signal achieves better ISAC performance -- i.e., higher sum throughput (up to $50\%$ peak throughput gain) for similar radar SNR (between $20$ to $24{\rm dB}$) -- than SDMA with a dedicated sensing signal. This first-ever experimental study of RSMA ISAC demonstrates the feasibility and the superiority of RSMA for future multi-functional wireless systems.

eess.SP↗

arXiv2025

RedOne: Revealing Domain-specific LLM Post-Training in Social Networking Services

As a primary medium for modern information dissemination, social networking services (SNS) have experienced rapid growth, which has proposed significant challenges for platform content management and interaction quality improvement. Recently, the development of large language models (LLMs) has offered potential solutions but existing studies focus on isolated tasks, which not only encounter diminishing benefit from the data scaling within individual scenarios but also fail to flexibly adapt to diverse real-world context. To address these challenges, we introduce RedOne, a domain-specific LLM designed to break the performance bottleneck of single-task baselines and establish a comprehensive foundation for the SNS. RedOne was developed through a three-stage training strategy consisting of continue pretraining, supervised fine-tuning, and preference optimization, using a large-scale real-world dataset. Through extensive experiments, RedOne maintains strong general capabilities, and achieves an average improvement up to 14.02% across 8 major SNS tasks and 7.56% in SNS bilingual evaluation benchmark, compared with base models. Furthermore, through online testing, RedOne reduced the exposure rate in harmful content detection by 11.23% and improved the click page rate in post-view search by 14.95% compared with single-tasks finetuned baseline models. These results establish RedOne as a robust domain-specific LLM for SNS, demonstrating excellent generalization across various tasks and promising applicability in real-world scenarios.

cs.LG↗

arXiv2025

Redefining Machine Translation on Social Network Services with Large Language Models

The globalization of social interactions has heightened the need for machine translation (MT) on Social Network Services (SNS), yet traditional models struggle with culturally nuanced content like memes, slang, and pop culture references. While large language models (LLMs) have advanced general-purpose translation, their performance on SNS-specific content remains limited due to insufficient specialized training data and evaluation benchmarks. This paper introduces RedTrans, a 72B LLM tailored for SNS translation, trained on a novel dataset developed through three innovations: (1) Supervised Finetuning with Dual-LLM Back-Translation Sampling, an unsupervised sampling method using LLM-based back-translation to select diverse data for large-scale finetuning; (2) Rewritten Preference Optimization (RePO), an algorithm that identifies and corrects erroneous preference pairs through expert annotation, building reliable preference corpora; and (3) RedTrans-Bench, the first benchmark for SNS translation, evaluating phenomena like humor localization, emoji semantics, and meme adaptation. Experiments show RedTrans outperforms state-of-the-art LLMs. Besides, RedTrans has already been deployed in a real-world production environment, demonstrating that domain-specific adaptation, effectively bridges the gap between generic and culturally grounded translation systems.

cs.CL↗

arXiv2025

Rate-Splitting Multiple Access for 6G: Prototypes, Experimental Results and Link/System level Simulations

Rate-Splitting Multiple Access (RSMA) is a powerful and versatile physical layer multiple access technique that generalizes and has better interference management capabilities than 5G-based Space Division Multiple Access (SDMA). It is also a rapidly maturing technology, all of which makes it a natural successor to SDMA in 6G. In this article, we describe RSMA's suitability for 6G by presenting: i) link and system level simulations of RSMA's performance gains over SDMA in realistic environments, and (ii) pioneering experimental results that demonstrate RSMA's gains over SDMA for key use cases like enhanced Mobile Broadband (eMBb), and Integrated Sensing and Communications (ISAC). We also comment on the status of standardization activities for RSMA.

eess.SP↗

arXiv2024

Rate-Splitting Multiple Access for Overloaded Multi-group Multicast: A First Experimental Study

Multi-group multicast (MGM) is an increasingly important form of multi-user wireless communications with several potential applications, such as video streaming, federated learning, safety-critical vehicular communications, etc. Rate-Splitting Multiple Access (RSMA) is a powerful interference management technique that can, in principle, achieve higher data rates and greater fairness for all types of multi-user wireless communications, including MGM. This paper presents the first-ever experimental evaluation of RSMA-based MGM, as well as the first-ever three-way comparison of RSMA-based, Space Divison Multiple Access (SDMA)-based and Non-Orthogonal Multiple Access (NOMA)-based MGM. Using a measurement setup involving a two-antenna transmitter and two groups of two single-antenna users per group, we consider the problem of realizing throughput (max-min) fairness across groups for each of three multiple access schemes, over nine experimental cases in a line-of-sight environment capturing varying levels of pathloss difference and channel correlation across the groups. Over these cases, we observe that RSMA-based MGM achieves fairness at a higher throughput for each group than SDMA- and NOMA-based MGM. These findings validate RSMA-based MGM's promised gains from the theoretical literature.

eess.SP↗

arXiv2024

Rate-Splitting Multiple Access for Non-Orthogonal Unicast Multicast: An Experimental Study

Non-orthogonal unicast multicast (NOUM) is a variant of multi-antenna multi-user communications where the users desire a shared message (multicast) in addition to their respective unique messages (unicast). The multicast rate is capped in many emerging NOUM applications, such as live-event broadcasting, location-based services and vehicular communications. Given this constraint, we experimentally show that when the user channels are highly correlated, Rate-Splitting Multiple Access (RSMA)-based NOUM can meet the multicast rate while supporting a larger unicast sum rate than multi-user linear precoding (MULP)-based NOUM and orthogonal unicast multicast.

eess.SP↗

arXiv2023

Rate-Splitting Multiple Access: The First Prototype and Experimental Validation of its Superiority over SDMA and NOMA

In multi-user multi-antenna communications, it is well-known in theory that Rate-Splitting Multiple Access (RSMA) can achieve a higher spectral efficiency than both Space Division Multiple Access (SDMA) and Non-Orthogonal Multiple Access (NOMA). However, an experimental evaluation of RSMA's performance, relative to SDMA and NOMA, is missing in the literature, which is essential to address the ongoing debate between RSMA and NOMA over which is better suited to handle most efficiently the available resources and interference in 6G. In this paper, we address this critical knowledge gap by realizing the first-ever RSMA prototype using software-defined radios. Through measurements using our prototype, we empirically solve the modulation and coding scheme limited sum throughput maximization problem for RSMA, SDMA and NOMA for the two-user multiple-input single-output (MISO) scenario over (a) different pairs of line-of-sight channels that vary in terms of their relative pathloss and spatial correlation, and with (b) different channel state information quality. We observe that RSMA achieves the highest sum throughput across all these cases, whereas SDMA and NOMA are effective only in some cases. Furthermore, RSMA also achieves better fairness at a higher sum throughput than both SDMA and NOMA.

eess.SP↗

arXiv2020

Rule-Guided Graph Neural Networks for Recommender Systems

To alleviate the cold start problem caused by collaborative filtering in recommender systems, knowledge graphs (KGs) are increasingly employed by many methods as auxiliary resources. However, existing work incorporated with KGs cannot capture the explicit long-range semantics between users and items meanwhile consider various connectivity between items. In this paper, we propose RGRec, which combines rule learning and graph neural networks (GNNs) for recommendation. RGRec first maps items to corresponding entities in KGs and adds users as new entities. Then, it automatically learns rules to model the explicit long-range semantics, and captures the connectivity between entities by aggregation to better encode various information. We show the effectiveness of RGRec on three real-world datasets. Particularly, the combination of rule learning and GNNs achieves substantial improvement compared to methods only using either of them.

cs.LG↗