Showing 1–35 of 35 results for author: Lai, G

Kimi Linear: An Expressive, Efficient Attention Architecture

Authors: Kimi Team, Yu Zhang, Zongyu Lin, Xingcheng Yao, Jiaxi Hu, Fanqing Meng, Chengyin Liu, Xin Men, Songlin Yang, Zhiyuan Li, Wentao Li, Enzhe Lu, Weizhou Liu, Yanru Chen, Weixin Xu, Longhui Yu, Yejie Wang, Yu Fan, Longguang Zhong, Enming Yuan, Dehao Zhang, Yizhi Zhang, T. Y. Liu, Haiming Wang, Shengjun Fang , et al. (35 additional authors not shown)

Abstract: We introduce Kimi Linear, a hybrid linear attention architecture that, for the first time, outperforms full attention under fair comparisons across various scenarios -- including short-context, long-context, and reinforcement learning (RL) scaling regimes. At its core lies Kimi Delta Attention (KDA), an expressive linear attention module that extends Gated DeltaNet with a finer-grained gating mech… ▽ More We introduce Kimi Linear, a hybrid linear attention architecture that, for the first time, outperforms full attention under fair comparisons across various scenarios -- including short-context, long-context, and reinforcement learning (RL) scaling regimes. At its core lies Kimi Delta Attention (KDA), an expressive linear attention module that extends Gated DeltaNet with a finer-grained gating mechanism, enabling more effective use of limited finite-state RNN memory. Our bespoke chunkwise algorithm achieves high hardware efficiency through a specialized variant of the Diagonal-Plus-Low-Rank (DPLR) transition matrices, which substantially reduces computation compared to the general DPLR formulation while remaining more consistent with the classical delta rule. We pretrain a Kimi Linear model with 3B activated parameters and 48B total parameters, based on a layerwise hybrid of KDA and Multi-Head Latent Attention (MLA). Our experiments show that with an identical training recipe, Kimi Linear outperforms full MLA with a sizeable margin across all evaluated tasks, while reducing KV cache usage by up to 75% and achieving up to 6 times decoding throughput for a 1M context. These results demonstrate that Kimi Linear can be a drop-in replacement for full attention architectures with superior performance and efficiency, including tasks with longer input and output lengths. To support further research, we open-source the KDA kernel and vLLM implementations, and release the pre-trained and instruction-tuned model checkpoints. △ Less

Submitted 1 November, 2025; v1 submitted 30 October, 2025; originally announced October 2025.

Comments: Kimi Linear tech report

The Lie of the Average: How Class Incremental Learning Evaluation Deceives You?

Authors: Guannan Lai, Da-Wei Zhou, Xin Yang, Han-Jia Ye

Abstract: Class Incremental Learning (CIL) requires models to continuously learn new classes without forgetting previously learned ones, while maintaining stable performance across all possible class sequences. In real-world settings, the order in which classes arrive is diverse and unpredictable, and model performance can vary substantially across different sequences. Yet mainstream evaluation protocols ca… ▽ More Class Incremental Learning (CIL) requires models to continuously learn new classes without forgetting previously learned ones, while maintaining stable performance across all possible class sequences. In real-world settings, the order in which classes arrive is diverse and unpredictable, and model performance can vary substantially across different sequences. Yet mainstream evaluation protocols calculate mean and variance from only a small set of randomly sampled sequences. Our theoretical analysis and empirical results demonstrate that this sampling strategy fails to capture the full performance range, resulting in biased mean estimates and a severe underestimation of the true variance in the performance distribution. We therefore contend that a robust CIL evaluation protocol should accurately characterize and estimate the entire performance distribution. To this end, we introduce the concept of extreme sequences and provide theoretical justification for their crucial role in the reliable evaluation of CIL. Moreover, we observe a consistent positive correlation between inter-task similarity and model performance, a relation that can be leveraged to guide the search for extreme sequences. Building on these insights, we propose EDGE (Extreme case-based Distribution and Generalization Evaluation), an evaluation protocol that adaptively identifies and samples extreme class sequences using inter-task similarity, offering a closer approximation of the ground-truth performance distribution. Extensive experiments demonstrate that EDGE effectively captures performance extremes and yields more accurate estimates of distributional boundaries, providing actionable insights for model selection and robustness checking. Our code is available at https://github.com/AIGNLAI/EDGE. △ Less

Submitted 26 September, 2025; originally announced September 2025.

Kimi K2: Open Agentic Intelligence

Authors: Kimi Team, Yifan Bai, Yiping Bao, Guanduo Chen, Jiahao Chen, Ningxin Chen, Ruijue Chen, Yanru Chen, Yuankun Chen, Yutian Chen, Zhuofu Chen, Jialei Cui, Hao Ding, Mengnan Dong, Angang Du, Chenzhuang Du, Dikang Du, Yulun Du, Yu Fan, Yichen Feng, Kelin Fu, Bofei Gao, Hongcheng Gao, Peizhong Gao, Tong Gao , et al. (144 additional authors not shown)

Abstract: We introduce Kimi K2, a Mixture-of-Experts (MoE) large language model with 32 billion activated parameters and 1 trillion total parameters. We propose the MuonClip optimizer, which improves upon Muon with a novel QK-clip technique to address training instability while enjoying the advanced token efficiency of Muon. Based on MuonClip, K2 was pre-trained on 15.5 trillion tokens with zero loss spike.… ▽ More We introduce Kimi K2, a Mixture-of-Experts (MoE) large language model with 32 billion activated parameters and 1 trillion total parameters. We propose the MuonClip optimizer, which improves upon Muon with a novel QK-clip technique to address training instability while enjoying the advanced token efficiency of Muon. Based on MuonClip, K2 was pre-trained on 15.5 trillion tokens with zero loss spike. During post-training, K2 undergoes a multi-stage post-training process, highlighted by a large-scale agentic data synthesis pipeline and a joint reinforcement learning (RL) stage, where the model improves its capabilities through interactions with real and synthetic environments. Kimi K2 achieves state-of-the-art performance among open-source non-thinking models, with strengths in agentic capabilities. Notably, K2 obtains 66.1 on Tau2-Bench, 76.5 on ACEBench (En), 65.8 on SWE-Bench Verified, and 47.3 on SWE-Bench Multilingual -- surpassing most open and closed-sourced baselines in non-thinking settings. It also exhibits strong capabilities in coding, mathematics, and reasoning tasks, with a score of 53.7 on LiveCodeBench v6, 49.5 on AIME 2025, 75.1 on GPQA-Diamond, and 27.1 on OJBench, all without extended thinking. These results position Kimi K2 as one of the most capable open-source large language models to date, particularly in software engineering and agentic tasks. We release our base and post-trained model checkpoints to facilitate future research and applications of agentic intelligence. △ Less

Submitted 28 July, 2025; originally announced July 2025.

Comments: tech report of Kimi K2

Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities

Authors: Gheorghe Comanici, Eric Bieber, Mike Schaekermann, Ice Pasupat, Noveen Sachdeva, Inderjit Dhillon, Marcel Blistein, Ori Ram, Dan Zhang, Evan Rosen, Luke Marris, Sam Petulla, Colin Gaffney, Asaf Aharoni, Nathan Lintz, Tiago Cardal Pais, Henrik Jacobsson, Idan Szpektor, Nan-Jiang Jiang, Krishna Haridasan, Ahmed Omran, Nikunj Saunshi, Dara Bahri, Gaurav Mishra, Eric Chu , et al. (3410 additional authors not shown)

Abstract: In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal unde… ▽ More In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal understanding and it is now able to process up to 3 hours of video content. Its unique combination of long context, multimodal and reasoning capabilities can be combined to unlock new agentic workflows. Gemini 2.5 Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements and Gemini 2.0 Flash and Flash-Lite provide high performance at low latency and cost. Taken together, the Gemini 2.X model generation spans the full Pareto frontier of model capability vs cost, allowing users to explore the boundaries of what is possible with complex agentic problem solving. △ Less

Submitted 16 October, 2025; v1 submitted 7 July, 2025; originally announced July 2025.

Comments: 72 pages, 17 figures

Kimi-Audio Technical Report

Authors: KimiTeam, Ding Ding, Zeqian Ju, Yichong Leng, Songxiang Liu, Tong Liu, Zeyu Shang, Kai Shen, Wei Song, Xu Tan, Heyi Tang, Zhengtao Wang, Chu Wei, Yifei Xin, Xinran Xu, Jianwei Yu, Yutao Zhang, Xinyu Zhou, Y. Charles, Jun Chen, Yanru Chen, Yulun Du, Weiran He, Zhenxing Hu, Guokun Lai , et al. (15 additional authors not shown)

Abstract: We present Kimi-Audio, an open-source audio foundation model that excels in audio understanding, generation, and conversation. We detail the practices in building Kimi-Audio, including model architecture, data curation, training recipe, inference deployment, and evaluation. Specifically, we leverage a 12.5Hz audio tokenizer, design a novel LLM-based architecture with continuous features as input a… ▽ More We present Kimi-Audio, an open-source audio foundation model that excels in audio understanding, generation, and conversation. We detail the practices in building Kimi-Audio, including model architecture, data curation, training recipe, inference deployment, and evaluation. Specifically, we leverage a 12.5Hz audio tokenizer, design a novel LLM-based architecture with continuous features as input and discrete tokens as output, and develop a chunk-wise streaming detokenizer based on flow matching. We curate a pre-training dataset that consists of more than 13 million hours of audio data covering a wide range of modalities including speech, sound, and music, and build a pipeline to construct high-quality and diverse post-training data. Initialized from a pre-trained LLM, Kimi-Audio is continual pre-trained on both audio and text data with several carefully designed tasks, and then fine-tuned to support a diverse of audio-related tasks. Extensive evaluation shows that Kimi-Audio achieves state-of-the-art performance on a range of audio benchmarks including speech recognition, audio understanding, audio question answering, and speech conversation. We release the codes, model checkpoints, as well as the evaluation toolkits in https://github.com/MoonshotAI/Kimi-Audio. △ Less

Submitted 25 April, 2025; originally announced April 2025.

Kimi-VL Technical Report

Authors: Kimi Team, Angang Du, Bohong Yin, Bowei Xing, Bowen Qu, Bowen Wang, Cheng Chen, Chenlin Zhang, Chenzhuang Du, Chu Wei, Congcong Wang, Dehao Zhang, Dikang Du, Dongliang Wang, Enming Yuan, Enzhe Lu, Fang Li, Flood Sung, Guangda Wei, Guokun Lai, Han Zhu, Hao Ding, Hao Hu, Hao Yang, Hao Zhang , et al. (70 additional authors not shown)

Abstract: We present Kimi-VL, an efficient open-source Mixture-of-Experts (MoE) vision-language model (VLM) that offers advanced multimodal reasoning, long-context understanding, and strong agent capabilities - all while activating only 2.8B parameters in its language decoder (Kimi-VL-A3B). Kimi-VL demonstrates strong performance across challenging domains: as a general-purpose VLM, Kimi-VL excels in multi-… ▽ More We present Kimi-VL, an efficient open-source Mixture-of-Experts (MoE) vision-language model (VLM) that offers advanced multimodal reasoning, long-context understanding, and strong agent capabilities - all while activating only 2.8B parameters in its language decoder (Kimi-VL-A3B). Kimi-VL demonstrates strong performance across challenging domains: as a general-purpose VLM, Kimi-VL excels in multi-turn agent tasks (e.g., OSWorld), matching flagship models. Furthermore, it exhibits remarkable capabilities across diverse challenging vision language tasks, including college-level image and video comprehension, OCR, mathematical reasoning, and multi-image understanding. In comparative evaluations, it effectively competes with cutting-edge efficient VLMs such as GPT-4o-mini, Qwen2.5-VL-7B, and Gemma-3-12B-IT, while surpassing GPT-4o in several key domains. Kimi-VL also advances in processing long contexts and perceiving clearly. With a 128K extended context window, Kimi-VL can process diverse long inputs, achieving impressive scores of 64.5 on LongVideoBench and 35.1 on MMLongBench-Doc. Its native-resolution vision encoder, MoonViT, further allows it to see and understand ultra-high-resolution visual inputs, achieving 83.2 on InfoVQA and 34.5 on ScreenSpot-Pro, while maintaining lower computational cost for common tasks. Building upon Kimi-VL, we introduce an advanced long-thinking variant: Kimi-VL-Thinking-2506. Developed through long chain-of-thought (CoT) supervised fine-tuning (SFT) and reinforcement learning (RL), the latest model exhibits strong long-horizon reasoning capabilities (64.0 on MMMU, 46.3 on MMMU-Pro, 56.9 on MathVision, 80.1 on MathVista, 65.2 on VideoMMMU) while obtaining robust general abilities. Code and models are publicly accessible at https://github.com/MoonshotAI/Kimi-VL. △ Less

Submitted 23 June, 2025; v1 submitted 10 April, 2025; originally announced April 2025.

Comments: Updated Kimi-VL-A3B-Thinking-2506 information

Iterative Prompting with Persuasion Skills in Jailbreaking Large Language Models

Authors: Shih-Wen Ke, Guan-Yu Lai, Guo-Lin Fang, Hsi-Yuan Kao

Abstract: Large language models (LLMs) are designed to align with human values in their responses. This study exploits LLMs with an iterative prompting technique where each prompt is systematically modified and refined across multiple iterations to enhance its effectiveness in jailbreaking attacks progressively. This technique involves analyzing the response patterns of LLMs, including GPT-3.5, GPT-4, LLaMa… ▽ More Large language models (LLMs) are designed to align with human values in their responses. This study exploits LLMs with an iterative prompting technique where each prompt is systematically modified and refined across multiple iterations to enhance its effectiveness in jailbreaking attacks progressively. This technique involves analyzing the response patterns of LLMs, including GPT-3.5, GPT-4, LLaMa2, Vicuna, and ChatGLM, allowing us to adjust and optimize prompts to evade the LLMs' ethical and security constraints. Persuasion strategies enhance prompt effectiveness while maintaining consistency with malicious intent. Our results show that the attack success rates (ASR) increase as the attacking prompts become more refined with the highest ASR of 90% for GPT4 and ChatGLM and the lowest ASR of 68% for LLaMa2. Our technique outperforms baseline techniques (PAIR and PAP) in ASR and shows comparable performance with GCG and ArtPrompt. △ Less

Submitted 26 March, 2025; originally announced March 2025.

Exploring Open-world Continual Learning with Knowns-Unknowns Knowledge Transfer

Authors: Yujie Li, Guannan Lai, Xin Yang, Yonghao Li, Marcello Bonsangue, Tianrui Li

Abstract: Open-World Continual Learning (OWCL) is a challenging paradigm where models must incrementally learn new knowledge without forgetting while operating under an open-world assumption. This requires handling incomplete training data and recognizing unknown samples during inference. However, existing OWCL methods often treat open detection and continual learning as separate tasks, limiting their abili… ▽ More Open-World Continual Learning (OWCL) is a challenging paradigm where models must incrementally learn new knowledge without forgetting while operating under an open-world assumption. This requires handling incomplete training data and recognizing unknown samples during inference. However, existing OWCL methods often treat open detection and continual learning as separate tasks, limiting their ability to integrate open-set detection and incremental classification in OWCL. Moreover, current approaches primarily focus on transferring knowledge from known samples, neglecting the insights derived from unknown/open samples. To address these limitations, we formalize four distinct OWCL scenarios and conduct comprehensive empirical experiments to explore potential challenges in OWCL. Our findings reveal a significant interplay between the open detection of unknowns and incremental classification of knowns, challenging a widely held assumption that unknown detection and known classification are orthogonal processes. Building on our insights, we propose \textbf{HoliTrans} (Holistic Knowns-Unknowns Knowledge Transfer), a novel OWCL framework that integrates nonlinear random projection (NRP) to create a more linearly separable embedding space and distribution-aware prototypes (DAPs) to construct an adaptive knowledge space. Particularly, our HoliTrans effectively supports knowledge transfer for both known and unknown samples while dynamically updating representations of open samples during OWCL. Extensive experiments across various OWCL scenarios demonstrate that HoliTrans outperforms 22 competitive baselines, bridging the gap between OWCL theory and practice and providing a robust, scalable framework for advancing open-world learning paradigms. △ Less

Submitted 27 February, 2025; originally announced February 2025.

Order-Robust Class Incremental Learning: Graph-Driven Dynamic Similarity Grouping

Authors: Guannan Lai, Yujie Li, Xiangkun Wang, Junbo Zhang, Tianrui Li, Xin Yang

Abstract: Class Incremental Learning (CIL) aims to enable models to learn new classes sequentially while retaining knowledge of previous ones. Although current methods have alleviated catastrophic forgetting (CF), recent studies highlight that the performance of CIL models is highly sensitive to the order of class arrival, particularly when sequentially introduced classes exhibit high inter-class similarity… ▽ More Class Incremental Learning (CIL) aims to enable models to learn new classes sequentially while retaining knowledge of previous ones. Although current methods have alleviated catastrophic forgetting (CF), recent studies highlight that the performance of CIL models is highly sensitive to the order of class arrival, particularly when sequentially introduced classes exhibit high inter-class similarity. To address this critical yet understudied challenge of class order sensitivity, we first extend existing CIL frameworks through theoretical analysis, proving that grouping classes with lower pairwise similarity during incremental phases significantly improves model robustness to order variations. Building on this insight, we propose Graph-Driven Dynamic Similarity Grouping (GDDSG), a novel method that employs graph coloring algorithms to dynamically partition classes into similarity-constrained groups. Each group trains an isolated CIL sub-model and constructs meta-features for class group identification. Experimental results demonstrate that our method effectively addresses the issue of class order sensitivity while achieving optimal performance in both model accuracy and anti-forgetting capability. Our code is available at https://github.com/AIGNLAI/GDDSG. △ Less

Submitted 17 March, 2025; v1 submitted 27 February, 2025; originally announced February 2025.

Comments: Accepted to CVPR 2025

MSC Class: 68T05; 68Q25; 68U05 ACM Class: I.2.6; I.2.10

Muon is Scalable for LLM Training

Authors: Jingyuan Liu, Jianlin Su, Xingcheng Yao, Zhejun Jiang, Guokun Lai, Yulun Du, Yidao Qin, Weixin Xu, Enzhe Lu, Junjie Yan, Yanru Chen, Huabin Zheng, Yibo Liu, Shaowei Liu, Bohong Yin, Weiran He, Han Zhu, Yuzhi Wang, Jianzhou Wang, Mengnan Dong, Zheng Zhang, Yongsheng Kang, Hao Zhang, Xinran Xu, Yutao Zhang , et al. (3 additional authors not shown)

Abstract: Recently, the Muon optimizer based on matrix orthogonalization has demonstrated strong results in training small-scale language models, but the scalability to larger models has not been proven. We identify two crucial techniques for scaling up Muon: (1) adding weight decay and (2) carefully adjusting the per-parameter update scale. These techniques allow Muon to work out-of-the-box on large-scale… ▽ More Recently, the Muon optimizer based on matrix orthogonalization has demonstrated strong results in training small-scale language models, but the scalability to larger models has not been proven. We identify two crucial techniques for scaling up Muon: (1) adding weight decay and (2) carefully adjusting the per-parameter update scale. These techniques allow Muon to work out-of-the-box on large-scale training without the need of hyper-parameter tuning. Scaling law experiments indicate that Muon achieves $\sim\!2\times$ computational efficiency compared to AdamW with compute optimal training. Based on these improvements, we introduce Moonlight, a 3B/16B-parameter Mixture-of-Expert (MoE) model trained with 5.7T tokens using Muon. Our model improves the current Pareto frontier, achieving better performance with much fewer training FLOPs compared to prior models. We open-source our distributed Muon implementation that is memory optimal and communication efficient. We also release the pretrained, instruction-tuned, and intermediate checkpoints to support future research. △ Less

Submitted 24 February, 2025; originally announced February 2025.

MoBA: Mixture of Block Attention for Long-Context LLMs

Authors: Enzhe Lu, Zhejun Jiang, Jingyuan Liu, Yulun Du, Tao Jiang, Chao Hong, Shaowei Liu, Weiran He, Enming Yuan, Yuzhi Wang, Zhiqi Huang, Huan Yuan, Suting Xu, Xinran Xu, Guokun Lai, Yanru Chen, Huabin Zheng, Junjie Yan, Jianlin Su, Yuxin Wu, Neo Y. Zhang, Zhilin Yang, Xinyu Zhou, Mingxing Zhang, Jiezhong Qiu

Abstract: Scaling the effective context length is essential for advancing large language models (LLMs) toward artificial general intelligence (AGI). However, the quadratic increase in computational complexity inherent in traditional attention mechanisms presents a prohibitive overhead. Existing approaches either impose strongly biased structures, such as sink or window attention which are task-specific, or… ▽ More Scaling the effective context length is essential for advancing large language models (LLMs) toward artificial general intelligence (AGI). However, the quadratic increase in computational complexity inherent in traditional attention mechanisms presents a prohibitive overhead. Existing approaches either impose strongly biased structures, such as sink or window attention which are task-specific, or radically modify the attention mechanism into linear approximations, whose performance in complex reasoning tasks remains inadequately explored. In this work, we propose a solution that adheres to the ``less structure'' principle, allowing the model to determine where to attend autonomously, rather than introducing predefined biases. We introduce Mixture of Block Attention (MoBA), an innovative approach that applies the principles of Mixture of Experts (MoE) to the attention mechanism. This novel architecture demonstrates superior performance on long-context tasks while offering a key advantage: the ability to seamlessly transition between full and sparse attention, enhancing efficiency without the risk of compromising performance. MoBA has already been deployed to support Kimi's long-context requests and demonstrates significant advancements in efficient attention computation for LLMs. Our code is available at https://github.com/MoonshotAI/MoBA. △ Less

Submitted 18 February, 2025; originally announced February 2025.

Comments: 15 pages

Kimi k1.5: Scaling Reinforcement Learning with LLMs

Authors: Kimi Team, Angang Du, Bofei Gao, Bowei Xing, Changjiu Jiang, Cheng Chen, Cheng Li, Chenjun Xiao, Chenzhuang Du, Chonghua Liao, Chuning Tang, Congcong Wang, Dehao Zhang, Enming Yuan, Enzhe Lu, Fengxiang Tang, Flood Sung, Guangda Wei, Guokun Lai, Haiqing Guo, Han Zhu, Hao Ding, Hao Hu, Hao Yang, Hao Zhang , et al. (71 additional authors not shown)

Abstract: Language model pretraining with next token prediction has proved effective for scaling compute but is limited to the amount of available training data. Scaling reinforcement learning (RL) unlocks a new axis for the continued improvement of artificial intelligence, with the promise that large language models (LLMs) can scale their training data by learning to explore with rewards. However, prior pu… ▽ More Language model pretraining with next token prediction has proved effective for scaling compute but is limited to the amount of available training data. Scaling reinforcement learning (RL) unlocks a new axis for the continued improvement of artificial intelligence, with the promise that large language models (LLMs) can scale their training data by learning to explore with rewards. However, prior published work has not produced competitive results. In light of this, we report on the training practice of Kimi k1.5, our latest multi-modal LLM trained with RL, including its RL training techniques, multi-modal data recipes, and infrastructure optimization. Long context scaling and improved policy optimization methods are key ingredients of our approach, which establishes a simplistic, effective RL framework without relying on more complex techniques such as Monte Carlo tree search, value functions, and process reward models. Notably, our system achieves state-of-the-art reasoning performance across multiple benchmarks and modalities -- e.g., 77.5 on AIME, 96.2 on MATH 500, 94-th percentile on Codeforces, 74.9 on MathVista -- matching OpenAI's o1. Moreover, we present effective long2short methods that use long-CoT techniques to improve short-CoT models, yielding state-of-the-art short-CoT reasoning results -- e.g., 60.8 on AIME, 94.6 on MATH500, 47.3 on LiveCodeBench -- outperforming existing short-CoT models such as GPT-4o and Claude Sonnet 3.5 by a large margin (up to +550%). △ Less

Submitted 2 June, 2025; v1 submitted 21 January, 2025; originally announced January 2025.

Comments: 25 pages

A New Perspective on Privacy Protection in Federated Learning with Granular-Ball Computing

Authors: Guannan Lai, Yihui Feng, Xin Yang, Xiaoyu Deng, Hao Yu, Shuyin Xia, Guoyin Wang, Tianrui Li

Abstract: Federated Learning (FL) facilitates collaborative model training while prioritizing privacy by avoiding direct data sharing. However, most existing articles attempt to address challenges within the model's internal parameters and corresponding outputs, while neglecting to solve them at the input level. To address this gap, we propose a novel framework called Granular-Ball Federated Learning (GrBFL… ▽ More Federated Learning (FL) facilitates collaborative model training while prioritizing privacy by avoiding direct data sharing. However, most existing articles attempt to address challenges within the model's internal parameters and corresponding outputs, while neglecting to solve them at the input level. To address this gap, we propose a novel framework called Granular-Ball Federated Learning (GrBFL) for image classification. GrBFL diverges from traditional methods that rely on the finest-grained input data. Instead, it segments images into multiple regions with optimal coarse granularity, which are then reconstructed into a graph structure. We designed a two-dimensional binary search segmentation algorithm based on variance constraints for GrBFL, which effectively removes redundant information while preserving key representative features. Extensive theoretical analysis and experiments demonstrate that GrBFL not only safeguards privacy and enhances efficiency but also maintains robust utility, consistently outperforming other state-of-the-art FL methods. The code is available at https://github.com/AIGNLAI/GrBFL. △ Less

Submitted 8 January, 2025; originally announced January 2025.

Creativity in the Age of AI: Evaluating the Impact of Generative AI on Design Outputs and Designers' Creative Thinking

Authors: Yue Fu, Han Bin, Tony Zhou, Marx Wang, Yixin Chen, Zelia Gomes Da Costa Lai, Jacob O. Wobbrock, Alexis Hiniker

Abstract: As generative AI (GenAI) increasingly permeates design workflows, its impact on design outcomes and designers' creative capabilities warrants investigation. We conducted a within-subjects experiment where we asked participants to design advertisements both with and without GenAI support. Our results show that expert evaluators rated GenAI-supported designs as more creative and unconventional ("wei… ▽ More As generative AI (GenAI) increasingly permeates design workflows, its impact on design outcomes and designers' creative capabilities warrants investigation. We conducted a within-subjects experiment where we asked participants to design advertisements both with and without GenAI support. Our results show that expert evaluators rated GenAI-supported designs as more creative and unconventional ("weird") despite no significant differences in visual appeal, brand alignment, or usefulness, which highlights the decoupling of novelty from usefulness-traditional dual components of creativity-in the context of GenAI usage. Moreover, while GenAI does not significantly enhance designers' overall creative thinking abilities, users were affected differently based on native language and prior AI exposure. Native English speakers experienced reduced relaxation when using AI, whereas designers new to GenAI exhibited gains in divergent thinking, such as idea fluency and flexibility. These findings underscore the variable impact of GenAI on different user groups, suggesting the potential for customized AI tools. △ Less

Submitted 31 October, 2024; originally announced November 2024.

Who is Undercover? Guiding LLMs to Explore Multi-Perspective Team Tactic in the Game

Authors: Ruiqi Dong, Zhixuan Liao, Guangwei Lai, Yuhan Ma, Danni Ma, Chenyou Fan

Abstract: Large Language Models (LLMs) are pivotal AI agents in complex tasks but still face challenges in open decision-making problems within complex scenarios. To address this, we use the language logic game ``Who is Undercover?'' (WIU) as an experimental platform to propose the Multi-Perspective Team Tactic (MPTT) framework. MPTT aims to cultivate LLMs' human-like language expression logic, multi-dimens… ▽ More Large Language Models (LLMs) are pivotal AI agents in complex tasks but still face challenges in open decision-making problems within complex scenarios. To address this, we use the language logic game ``Who is Undercover?'' (WIU) as an experimental platform to propose the Multi-Perspective Team Tactic (MPTT) framework. MPTT aims to cultivate LLMs' human-like language expression logic, multi-dimensional thinking, and self-perception in complex scenarios. By alternating speaking and voting sessions, integrating techniques like self-perspective, identity-determination, self-reflection, self-summary and multi-round find-teammates, LLM agents make rational decisions through strategic concealment and communication, fostering human-like trust. Preliminary results show that MPTT, combined with WIU, leverages LLMs' cognitive capabilities to create a decision-making framework that can simulate real society. This framework aids minority groups in communication and expression, promoting fairness and diversity in decision-making. Additionally, our Human-in-the-loop experiments demonstrate that LLMs can learn and align with human behaviors through interactive, indicating their potential for active participation in societal decision-making. △ Less

Submitted 20 October, 2024; originally announced October 2024.

Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

Authors: Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent, Zhufeng Pan, Shibo Wang, Soroosh Mariooryad, Yifan Ding, Xinyang Geng, Fred Alcober, Roy Frostig, Mark Omernick, Lexi Walker, Cosmin Paduraru, Christina Sorokin, Andrea Tacchetti, Colin Gaffney, Samira Daruki, Olcan Sercinoglu, Zach Gleicher, Juliette Love , et al. (1112 additional authors not shown)

Abstract: In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February… ▽ More In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content. △ Less

Submitted 16 December, 2024; v1 submitted 8 March, 2024; originally announced March 2024.

Should ChatGPT Write Your Breakup Text? Exploring the Role of AI in Relationship Dissolution

Authors: Yue Fu, Yixin Chen, Zelia Gomes Da Costa Lai, Alexis Hiniker

Abstract: Relationships are essential to our happiness and wellbeing, yet their dissolution-the final stage of a relationship's lifecycle-is among the most stressful events individuals can experience, often leading to profound and lasting impacts. With the breakup process increasingly facilitated by technology, such as computer-mediated communication, and the likely future influence of generative AI (GenAI)… ▽ More Relationships are essential to our happiness and wellbeing, yet their dissolution-the final stage of a relationship's lifecycle-is among the most stressful events individuals can experience, often leading to profound and lasting impacts. With the breakup process increasingly facilitated by technology, such as computer-mediated communication, and the likely future influence of generative AI (GenAI) tools, we conducted a semi-structured interview study with 21 participants. We aim to understand: 1) the current role of technology in the breakup process, 2) the needs and support individuals seek during this time, and 3) how GenAI might address or undermine these needs. Our findings show that people have distinct needs at various stages of breakups. While currently technology plays an important role, it falls short in supporting users' unmet needs. Participants envision that GenAI could: 1) aid in prompting self-reflection, providing neutral second opinions, and assisting with planning leading up to a breakup; 2) serve as a communication mediator, supporting wording and tone to facilitate emotional expression during breakup conversations; and 3) support personal growth and offer companionship after a breakup. However, our findings also reveal participants' concerns about involving GenAI in this process. Based on our results, we discuss the potential opportunities, design considerations, and harms of GenAI tools in facilitating people's relationship dissolution. △ Less

Submitted 31 October, 2024; v1 submitted 17 January, 2024; originally announced January 2024.

Report number: 2025 Computer-Supported Cooperative Work & Social Computing (CSCW)

A Self-enhancement Approach for Domain-specific Chatbot Training via Knowledge Mining and Digest

Authors: Ruohong Zhang, Luyu Gao, Chen Zheng, Zhen Fan, Guokun Lai, Zheng Zhang, Fangzhou Ai, Yiming Yang, Hongxia Yang

Abstract: Large Language Models (LLMs), despite their great power in language generation, often encounter challenges when dealing with intricate and knowledge-demanding queries in specific domains. This paper introduces a novel approach to enhance LLMs by effectively extracting the relevant knowledge from domain-specific textual sources, and the adaptive training of a chatbot with domain-specific inquiries.… ▽ More Large Language Models (LLMs), despite their great power in language generation, often encounter challenges when dealing with intricate and knowledge-demanding queries in specific domains. This paper introduces a novel approach to enhance LLMs by effectively extracting the relevant knowledge from domain-specific textual sources, and the adaptive training of a chatbot with domain-specific inquiries. Our two-step approach starts from training a knowledge miner, namely LLMiner, which autonomously extracts Question-Answer pairs from relevant documents through a chain-of-thought reasoning process. Subsequently, we blend the mined QA pairs with a conversational dataset to fine-tune the LLM as a chatbot, thereby enriching its domain-specific expertise and conversational capabilities. We also developed a new evaluation benchmark which comprises four domain-specific text corpora and associated human-crafted QA pairs for testing. Our model shows remarkable performance improvement over generally aligned LLM and surpasses domain-adapted models directly fine-tuned on domain corpus. In particular, LLMiner achieves this with minimal human intervention, requiring only 600 seed instances, thereby providing a pathway towards self-improvement of LLMs through model-synthesized training data. △ Less

Submitted 17 November, 2023; originally announced November 2023.

Comments: Work in progress

N-Grammer: Augmenting Transformers with latent n-grams

Authors: Aurko Roy, Rohan Anil, Guangda Lai, Benjamin Lee, Jeffrey Zhao, Shuyuan Zhang, Shibo Wang, Ye Zhang, Shen Wu, Rigel Swavely, Tao, Yu, Phuong Dao, Christopher Fifty, Zhifeng Chen, Yonghui Wu

Abstract: Transformer models have recently emerged as one of the foundational models in natural language processing, and as a byproduct, there is significant recent interest and investment in scaling these models. However, the training and inference costs of these large Transformer language models are prohibitive, thus necessitating more research in identifying more efficient variants. In this work, we prop… ▽ More Transformer models have recently emerged as one of the foundational models in natural language processing, and as a byproduct, there is significant recent interest and investment in scaling these models. However, the training and inference costs of these large Transformer language models are prohibitive, thus necessitating more research in identifying more efficient variants. In this work, we propose a simple yet effective modification to the Transformer architecture inspired by the literature in statistical language modeling, by augmenting the model with n-grams that are constructed from a discrete latent representation of the text sequence. We evaluate our model, the N-Grammer on language modeling on the C4 data-set as well as text classification on the SuperGLUE data-set, and find that it outperforms several strong baselines such as the Transformer and the Primer. We open-source our model for reproducibility purposes in Jax. △ Less

Submitted 13 July, 2022; originally announced July 2022.

Comments: 8 pages, 2 figures

Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks

Authors: Yizhong Wang, Swaroop Mishra, Pegah Alipoormolabashi, Yeganeh Kordi, Amirreza Mirzaei, Anjana Arunkumar, Arjun Ashok, Arut Selvan Dhanasekaran, Atharva Naik, David Stap, Eshaan Pathak, Giannis Karamanolakis, Haizhi Gary Lai, Ishan Purohit, Ishani Mondal, Jacob Anderson, Kirby Kuznia, Krima Doshi, Maitreya Patel, Kuntal Kumar Pal, Mehrad Moradshahi, Mihir Parmar, Mirali Purohit, Neeraj Varshney, Phani Rohitha Kaza , et al. (15 additional authors not shown)

Abstract: How well can NLP models generalize to a variety of unseen tasks when provided with task instructions? To address this question, we first introduce Super-NaturalInstructions, a benchmark of 1,616 diverse NLP tasks and their expert-written instructions. Our collection covers 76 distinct task types, including but not limited to classification, extraction, infilling, sequence tagging, text rewriting,… ▽ More How well can NLP models generalize to a variety of unseen tasks when provided with task instructions? To address this question, we first introduce Super-NaturalInstructions, a benchmark of 1,616 diverse NLP tasks and their expert-written instructions. Our collection covers 76 distinct task types, including but not limited to classification, extraction, infilling, sequence tagging, text rewriting, and text composition. This large and diverse collection of tasks enables rigorous benchmarking of cross-task generalization under instructions -- training models to follow instructions on a subset of tasks and evaluating them on the remaining unseen ones. Furthermore, we build Tk-Instruct, a transformer model trained to follow a variety of in-context instructions (plain language task definitions or k-shot examples). Our experiments show that Tk-Instruct outperforms existing instruction-following models such as InstructGPT by over 9% on our benchmark despite being an order of magnitude smaller. We further analyze generalization as a function of various scaling parameters, such as the number of observed tasks, the number of instances per task, and model sizes. We hope our dataset and model facilitate future progress towards more general-purpose NLP models. △ Less

Submitted 24 October, 2022; v1 submitted 15 April, 2022; originally announced April 2022.

Comments: Accepted to EMNLP 2022, 25 pages

Interactive Evolutionary Multi-Objective Optimization via Learning-to-Rank

Authors: Ke Li, Guiyu Lai, Xin Yao

Abstract: In practical multi-criterion decision-making, it is cumbersome if a decision maker (DM) is asked to choose among a set of trade-off alternatives covering the whole Pareto-optimal front. This is a paradox in conventional evolutionary multi-objective optimization (EMO) that always aim to achieve a well balance between convergence and diversity. In essence, the ultimate goal of multi-objective optimi… ▽ More In practical multi-criterion decision-making, it is cumbersome if a decision maker (DM) is asked to choose among a set of trade-off alternatives covering the whole Pareto-optimal front. This is a paradox in conventional evolutionary multi-objective optimization (EMO) that always aim to achieve a well balance between convergence and diversity. In essence, the ultimate goal of multi-objective optimization is to help a decision maker (DM) identify solution(s) of interest (SOI) achieving satisfactory trade-offs among multiple conflicting criteria. Bearing this in mind, this paper develops a framework for designing preference-based EMO algorithms to find SOI in an interactive manner. Its core idea is to involve human in the loop of EMO. After every several iterations, the DM is invited to elicit her feedback with regard to a couple of incumbent candidates. By collecting such information, her preference is progressively learned by a learning-to-rank neural network and then applied to guide the baseline EMO algorithm. Note that this framework is so general that any existing EMO algorithm can be applied in a plug-in manner. Experiments on $48$ benchmark test problems with up to 10 objectives fully demonstrate the effectiveness of our proposed algorithms for finding SOI. △ Less

Submitted 6 April, 2022; originally announced April 2022.

Two Decades of Game Jams

Authors: Gorm Lai, Annakaisa Kultima, Foaad Khosmood, Johanna Pirker, Allan Fowler, Ilaria Vecchi, William Latham, Frederic Fol Leymarie

Abstract: In less than a year's time, March 2022 will mark the twentieth anniversary of the first documented game jam, the Indie Game Jam, which took place in Oakland, California in 2002. Initially, game jams were widely seen as frivolous activities. Since then, they have taken the world by storm. Game jams have not only become part of the day-to-day process of many game developers, but jams are also used f… ▽ More In less than a year's time, March 2022 will mark the twentieth anniversary of the first documented game jam, the Indie Game Jam, which took place in Oakland, California in 2002. Initially, game jams were widely seen as frivolous activities. Since then, they have taken the world by storm. Game jams have not only become part of the day-to-day process of many game developers, but jams are also used for activist purposes, for learning and teaching, as part of the experience economy, for making commercial prototypes that gamers can vote on, and more. Beyond only surveying game jams and the relevant published scientific literature from the last two decades, this paper has several additional contributions. It builds a history of game jams, and proposes two different taxonomies of game jams - a historical and a categorical. In addition, it discusses the definition of game jam and identifies the most active research areas within the game jam community such as the interplay and development with local communities, the study and analysis of game jammers and organisers, and works that bring a critical look on game jams. △ Less

Submitted 26 October, 2021; originally announced October 2021.

Journal ref: ICGJ 2021: Sixth Annual International Conference on Game Jams, Hackathons, and Game Creation Events

Unsupervised Parallel Corpus Mining on Web Data

Authors: Guokun Lai, Zihang Dai, Yiming Yang

Abstract: With a large amount of parallel data, neural machine translation systems are able to deliver human-level performance for sentence-level translation. However, it is costly to label a large amount of parallel data by humans. In contrast, there is a large-scale of parallel corpus created by humans on the Internet. The major difficulty to utilize them is how to filter them out from the noise website e… ▽ More With a large amount of parallel data, neural machine translation systems are able to deliver human-level performance for sentence-level translation. However, it is costly to label a large amount of parallel data by humans. In contrast, there is a large-scale of parallel corpus created by humans on the Internet. The major difficulty to utilize them is how to filter them out from the noise website environments. Current parallel data mining methods all require labeled parallel data as the training source. In this paper, we present a pipeline to mine the parallel corpus from the Internet in an unsupervised manner. On the widely used WMT'14 English-French and WMT'16 English-German benchmarks, the machine translator trained with the data extracted by our pipeline achieves very close performance to the supervised results. On the WMT'16 English-Romanian and Romanian-English benchmarks, our system produces new state-of-the-art results, 39.81 and 38.95 BLEU scores, even compared with supervised approaches. △ Less

Submitted 17 September, 2020; originally announced September 2020.

Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing

Authors: Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le

Abstract: With the success of language pretraining, it is highly desirable to develop more efficient architectures of good scalability that can exploit the abundant unlabeled data at a lower cost. To improve the efficiency, we examine the much-overlooked redundancy in maintaining a full-length token-level presentation, especially for tasks that only require a single-vector presentation of the sequence. With… ▽ More With the success of language pretraining, it is highly desirable to develop more efficient architectures of good scalability that can exploit the abundant unlabeled data at a lower cost. To improve the efficiency, we examine the much-overlooked redundancy in maintaining a full-length token-level presentation, especially for tasks that only require a single-vector presentation of the sequence. With this intuition, we propose Funnel-Transformer which gradually compresses the sequence of hidden states to a shorter one and hence reduces the computation cost. More importantly, by re-investing the saved FLOPs from length reduction in constructing a deeper or wider model, we further improve the model capacity. In addition, to perform token-level predictions as required by common pretraining objectives, Funnel-Transformer is able to recover a deep representation for each token from the reduced hidden sequence via a decoder. Empirically, with comparable or fewer FLOPs, Funnel-Transformer outperforms the standard Transformer on a wide variety of sequence-level prediction tasks, including text classification, language understanding, and reading comprehension. The code and pretrained checkpoints are available at https://github.com/laiguokun/Funnel-Transformer. △ Less

Submitted 5 June, 2020; originally announced June 2020.

Towards Friendly Mixed Initiative Procedural Content Generation: Three Pillars of Industry

Authors: Gorm Lai, William Latham, Frederic Fol Leymarie

Abstract: While the games industry is moving towards procedural content generation (PCG) with tools available under popular platforms such as Unreal, Unity or Houdini, and video game titles like No Man's Sky and Horizon Zero Dawn taking advantage of PCG, the gap between academia and industry is as wide as it has ever been, in terms of communication and sharing methods. One of the authors, has worked on both… ▽ More While the games industry is moving towards procedural content generation (PCG) with tools available under popular platforms such as Unreal, Unity or Houdini, and video game titles like No Man's Sky and Horizon Zero Dawn taking advantage of PCG, the gap between academia and industry is as wide as it has ever been, in terms of communication and sharing methods. One of the authors, has worked on both sides of this gap and in an effort to shorten it and increase the synergy between the two sectors, has identified three design pillars for PCG using mixed-initiative interfaces. The three pillars are Respect Designer Control, Respect the Creative Process and Respect Existing Work Processes. Respecting designer control is about creating a tool that gives enough control to bring out the designer's vision. Respecting the creative process concerns itself with having a feedback loop that is short enough, that the creative process is not disturbed. Respecting existing work processes means that a PCG tool should plug in easily to existing asset pipelines. As academics and communicators, it is surprising that publications often do not describe ways for developers to use our work or lack considerations for how a piece of work might fit into existing content pipelines. △ Less

Submitted 19 May, 2020; originally announced May 2020.

Correlation-aware Unsupervised Change-point Detection via Graph Neural Networks

Authors: Ruohong Zhang, Yu Hao, Donghan Yu, Wei-Cheng Chang, Guokun Lai, Yiming Yang

Abstract: Change-point detection (CPD) aims to detect abrupt changes over time series data. Intuitively, effective CPD over multivariate time series should require explicit modeling of the dependencies across input variables. However, existing CPD methods either ignore the dependency structures entirely or rely on the (unrealistic) assumption that the correlation structures are static over time. In this pap… ▽ More Change-point detection (CPD) aims to detect abrupt changes over time series data. Intuitively, effective CPD over multivariate time series should require explicit modeling of the dependencies across input variables. However, existing CPD methods either ignore the dependency structures entirely or rely on the (unrealistic) assumption that the correlation structures are static over time. In this paper, we propose a Correlation-aware Dynamics Model for CPD, which explicitly models the correlation structure and dynamics of variables by incorporating graph neural networks into an encoder-decoder framework. Extensive experiments on synthetic and real-world datasets demonstrate the advantageous performance of the proposed model on CPD tasks over strong baselines, as well as its ability to classify the change-points as correlation changes or independent changes. Keywords: Multivariate Time Series, Change-point Detection, Graph Neural Networks △ Less

Submitted 13 September, 2020; v1 submitted 24 April, 2020; originally announced April 2020.

Comments: Accepted for publication in the International Conference on Neural Information Processing (ICONIP) 2020 Original paper is 12 pages, additional appendix is available on arxiv

MSC Class: I.2.6

Journal ref: ICONIP 2020: Neural Information Processing

DOPS: Learning to Detect 3D Objects and Predict their 3D Shapes

Authors: Mahyar Najibi, Guangda Lai, Abhijit Kundu, Zhichao Lu, Vivek Rathod, Thomas Funkhouser, Caroline Pantofaru, David Ross, Larry S. Davis, Alireza Fathi

Abstract: We propose DOPS, a fast single-stage 3D object detection method for LIDAR data. Previous methods often make domain-specific design decisions, for example projecting points into a bird-eye view image in autonomous driving scenarios. In contrast, we propose a general-purpose method that works on both indoor and outdoor scenes. The core novelty of our method is a fast, single-pass architecture that b… ▽ More We propose DOPS, a fast single-stage 3D object detection method for LIDAR data. Previous methods often make domain-specific design decisions, for example projecting points into a bird-eye view image in autonomous driving scenarios. In contrast, we propose a general-purpose method that works on both indoor and outdoor scenes. The core novelty of our method is a fast, single-pass architecture that both detects objects in 3D and estimates their shapes. 3D bounding box parameters are estimated in one pass for every point, aggregated through graph convolutions, and fed into a branch of the network that predicts latent codes representing the shape of each detected object. The latent shape space and shape decoder are learned on a synthetic dataset and then used as supervision for the end-to-end training of the 3D object detection pipeline. Thus our model is able to extract shapes without access to ground-truth shape information in the target dataset. During experiments, we find that our proposed method achieves state-of-the-art results by ~5% on object detection in ScanNet scenes, and it gets top results by 3.4% in the Waymo Open Dataset, while reproducing the shapes of detected cars. △ Less

Submitted 6 April, 2020; v1 submitted 2 April, 2020; originally announced April 2020.

Comments: To appear in CVPR 2020

Bridging the domain gap in cross-lingual document classification

Authors: Guokun Lai, Barlas Oguz, Yiming Yang, Veselin Stoyanov

Abstract: The scarcity of labeled training data often prohibits the internationalization of NLP models to multiple languages. Recent developments in cross-lingual understanding (XLU) has made progress in this area, trying to bridge the language barrier using language universal representations. However, even if the language problem was resolved, models trained in one language would not transfer to another la… ▽ More The scarcity of labeled training data often prohibits the internationalization of NLP models to multiple languages. Recent developments in cross-lingual understanding (XLU) has made progress in this area, trying to bridge the language barrier using language universal representations. However, even if the language problem was resolved, models trained in one language would not transfer to another language perfectly due to the natural domain drift across languages and cultures. We consider the setting of semi-supervised cross-lingual understanding, where labeled data is available in a source language (English), but only unlabeled data is available in the target language. We combine state-of-the-art cross-lingual methods with recently proposed methods for weakly supervised learning such as unsupervised pre-training and unsupervised data augmentation to simultaneously close both the language gap and the domain gap in XLU. We show that addressing the domain gap is crucial. We improve over strong baselines and achieve a new state-of-the-art for cross-lingual document classification. △ Less

Submitted 20 September, 2019; v1 submitted 16 September, 2019; originally announced September 2019.

Introducing: The Game Jam License

Authors: Gorm Lai, Kai Erenli, Foaad Khosmood, William Latham

Abstract: Since their inception at the Indie Game Jam in 2002, a significant part of game jams has been knowledge sharing and showcasing ideas and work to peers. While various licensing mechanisms have been used for game jams throughout the years, there has never been a licence uniquely designed for artifacts created during a game jam. In this paper, we present to the community the Game Jam License (GJL) wh… ▽ More Since their inception at the Indie Game Jam in 2002, a significant part of game jams has been knowledge sharing and showcasing ideas and work to peers. While various licensing mechanisms have been used for game jams throughout the years, there has never been a licence uniquely designed for artifacts created during a game jam. In this paper, we present to the community the Game Jam License (GJL) which is designed to facilitate that sharing and knowledge transfer, while making sure the original creators retain commercial rights. The Global Game Jam, since 2009, strives to formalise sharing in a similar manner, by having jammers upload and license their creations under Creative Commons Non Commercial Share Alike 3.0 free license. However, the CC family of licenses is not well suited for software. CC is not compatible with most other licenses, and introduces a legal grey area with the division between commercial and non-commercial use. Moreover, open source licences like GPL are well suited for source code, but not for art and design content. Instead the GJL presented in this paper, aims to uphold the original ideas of game jams (sharing and knowledge transfer), while still allowing the original team to hold on to all rights to their creation, without any of the deficiencies of the CC family of licenses. △ Less

Submitted 29 August, 2019; originally announced August 2019.

Re-examination of the Role of Latent Variables in Sequence Modeling

Authors: Zihang Dai, Guokun Lai, Yiming Yang, Shinjae Yoo

Abstract: With latent variables, stochastic recurrent models have achieved state-of-the-art performance in modeling sound-wave sequence. However, opposite results are also observed in other domains, where standard recurrent networks often outperform stochastic models. To better understand this discrepancy, we re-examine the roles of latent variables in stochastic recurrent models for speech density estimati… ▽ More With latent variables, stochastic recurrent models have achieved state-of-the-art performance in modeling sound-wave sequence. However, opposite results are also observed in other domains, where standard recurrent networks often outperform stochastic models. To better understand this discrepancy, we re-examine the roles of latent variables in stochastic recurrent models for speech density estimation. Our analysis reveals that under the restriction of fully factorized output distribution in previous evaluations, the stochastic models were implicitly leveraging intra-step correlation but the standard recurrent baselines were prohibited to do so, resulting in an unfair comparison. To correct the unfairness, we remove such restriction in our re-examination, where all the models can explicitly leverage intra-step correlation with an auto-regressive structure. Over a diverse set of sequential data, including human speech, MIDI music, handwriting trajectory and frame-permuted speech, our results show that stochastic recurrent models fail to exhibit any practical advantage despite the claimed theoretical superiority. In contrast, standard recurrent models equipped with an auto-regressive output distribution consistently perform better, significantly advancing the state-of-the-art results on three speech datasets. △ Less

Submitted 16 September, 2019; v1 submitted 4 February, 2019; originally announced February 2019.

Comments: Code available at https://github.com/zihangdai/reexamine-srnn, accepted by NeurIPS 2019

Stochastic WaveNet: A Generative Latent Variable Model for Sequential Data

Authors: Guokun Lai, Bohan Li, Guoqing Zheng, Yiming Yang

Abstract: How to model distribution of sequential data, including but not limited to speech and human motions, is an important ongoing research problem. It has been demonstrated that model capacity can be significantly enhanced by introducing stochastic latent variables in the hidden states of recurrent neural networks. Simultaneously, WaveNet, equipped with dilated convolutions, achieves astonishing empiri… ▽ More How to model distribution of sequential data, including but not limited to speech and human motions, is an important ongoing research problem. It has been demonstrated that model capacity can be significantly enhanced by introducing stochastic latent variables in the hidden states of recurrent neural networks. Simultaneously, WaveNet, equipped with dilated convolutions, achieves astonishing empirical performance in natural speech generation task. In this paper, we combine the ideas from both stochastic latent variables and dilated convolutions, and propose a new architecture to model sequential data, termed as Stochastic WaveNet, where stochastic latent variables are injected into the WaveNet structure. We argue that Stochastic WaveNet enjoys powerful distribution modeling capacity and the advantage of parallel training from dilated convolutions. In order to efficiently infer the posterior distribution of the latent variables, a novel inference network structure is designed based on the characteristics of WaveNet architecture. State-of-the-art performances on benchmark datasets are obtained by Stochastic WaveNet on natural speech modeling and high quality human handwriting samples can be generated as well. △ Less

Submitted 15 June, 2018; originally announced June 2018.

Comments: ICML 2018 Workshop

Large-scale Cloze Test Dataset Created by Teachers

Authors: Qizhe Xie, Guokun Lai, Zihang Dai, Eduard Hovy

Abstract: Cloze tests are widely adopted in language exams to evaluate students' language proficiency. In this paper, we propose the first large-scale human-created cloze test dataset CLOTH, containing questions used in middle-school and high-school language exams. With missing blanks carefully created by teachers and candidate choices purposely designed to be nuanced, CLOTH requires a deeper language under… ▽ More Cloze tests are widely adopted in language exams to evaluate students' language proficiency. In this paper, we propose the first large-scale human-created cloze test dataset CLOTH, containing questions used in middle-school and high-school language exams. With missing blanks carefully created by teachers and candidate choices purposely designed to be nuanced, CLOTH requires a deeper language understanding and a wider attention span than previously automatically-generated cloze datasets. We test the performance of dedicatedly designed baseline models including a language model trained on the One Billion Word Corpus and show humans outperform them by a significant margin. We investigate the source of the performance gap, trace model deficiencies to some distinct properties of CLOTH, and identify the limited ability of comprehending the long-term context to be the key bottleneck. △ Less

Submitted 27 August, 2018; v1 submitted 8 November, 2017; originally announced November 2017.

Comments: EMNLP 2018

Learning Depthwise Separable Graph Convolution from Data Manifold

Authors: Guokun Lai, Hanxiao Liu, Yiming Yang

Abstract: Convolution Neural Network (CNN) has gained tremendous success in computer vision tasks with its outstanding ability to capture the local latent features. Recently, there has been an increasing interest in extending convolution operations to the non-Euclidean geometry. Although various types of convolution operations have been proposed for graphs or manifolds, their connections with traditional co… ▽ More Convolution Neural Network (CNN) has gained tremendous success in computer vision tasks with its outstanding ability to capture the local latent features. Recently, there has been an increasing interest in extending convolution operations to the non-Euclidean geometry. Although various types of convolution operations have been proposed for graphs or manifolds, their connections with traditional convolution over grid-structured data are not well-understood. In this paper, we show that depthwise separable convolution can be successfully generalized for the unification of both graph-based and grid-based convolution methods. Based on this insight we propose a novel Depthwise Separable Graph Convolution (DSGC) approach which is compatible with the tradition convolution network and subsumes existing convolution methods as special cases. It is equipped with the combined strengths in model expressiveness, compatibility (relatively small number of parameters), modularity and computational efficiency in training. Extensive experiments show the outstanding performance of DSGC in comparison with strong baselines on multi-domain benchmark datasets. △ Less

Submitted 8 November, 2018; v1 submitted 31 October, 2017; originally announced October 2017.

RACE: Large-scale ReAding Comprehension Dataset From Examinations

Authors: Guokun Lai, Qizhe Xie, Hanxiao Liu, Yiming Yang, Eduard Hovy

Abstract: We present RACE, a new dataset for benchmark evaluation of methods in the reading comprehension task. Collected from the English exams for middle and high school Chinese students in the age range between 12 to 18, RACE consists of near 28,000 passages and near 100,000 questions generated by human experts (English instructors), and covers a variety of topics which are carefully designed for evaluat… ▽ More We present RACE, a new dataset for benchmark evaluation of methods in the reading comprehension task. Collected from the English exams for middle and high school Chinese students in the age range between 12 to 18, RACE consists of near 28,000 passages and near 100,000 questions generated by human experts (English instructors), and covers a variety of topics which are carefully designed for evaluating the students' ability in understanding and reasoning. In particular, the proportion of questions that requires reasoning is much larger in RACE than that in other benchmark datasets for reading comprehension, and there is a significant gap between the performance of the state-of-the-art models (43%) and the ceiling human performance (95%). We hope this new dataset can serve as a valuable resource for research and evaluation in machine comprehension. The dataset is freely available at http://www.cs.cmu.edu/~glai1/data/race/ and the code is available at https://github.com/qizhex/RACE_AR_baselines. △ Less

Submitted 5 December, 2017; v1 submitted 15 April, 2017; originally announced April 2017.

Comments: EMNLP 2017

Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks

Authors: Guokun Lai, Wei-Cheng Chang, Yiming Yang, Hanxiao Liu

Abstract: Multivariate time series forecasting is an important machine learning problem across many domains, including predictions of solar plant energy output, electricity consumption, and traffic jam situation. Temporal data arise in these real-world applications often involves a mixture of long-term and short-term patterns, for which traditional approaches such as Autoregressive models and Gaussian Proce… ▽ More Multivariate time series forecasting is an important machine learning problem across many domains, including predictions of solar plant energy output, electricity consumption, and traffic jam situation. Temporal data arise in these real-world applications often involves a mixture of long-term and short-term patterns, for which traditional approaches such as Autoregressive models and Gaussian Process may fail. In this paper, we proposed a novel deep learning framework, namely Long- and Short-term Time-series network (LSTNet), to address this open challenge. LSTNet uses the Convolution Neural Network (CNN) and the Recurrent Neural Network (RNN) to extract short-term local dependency patterns among variables and to discover long-term patterns for time series trends. Furthermore, we leverage traditional autoregressive model to tackle the scale insensitive problem of the neural network model. In our evaluation on real-world data with complex mixtures of repetitive patterns, LSTNet achieved significant performance improvements over that of several state-of-the-art baseline methods. All the data and experiment codes are available online. △ Less

Submitted 18 April, 2018; v1 submitted 20 March, 2017; originally announced March 2017.

Comments: Accepted by SIGIR 2018