Xiao-Yang Liu

Xiao-Yang Liu, Ziyi Xia, Hongyang Yang et al.

The financial market is a particularly challenging playground for deep reinforcement learning due to its unique feature of dynamic datasets. Building high-quality market environments for training financial reinforcement learning (FinRL) agents is difficult due to major factors such as the low signal-to-noise ratio of financial data, survivorship bias of historical data, and model overfitting. In this paper, we present FinRL-Meta, a data-centric and openly accessible library that processes dynamic datasets from real-world markets into gym-style market environments and has been actively maintained by the AI4Finance community. First, following a DataOps paradigm, we provide hundreds of market environments through an automatic data curation pipeline. Second, we provide homegrown examples and reproduce popular research papers as stepping stones for users to design new trading strategies. We also deploy the library on cloud platforms so that users can visualize their own results and assess the relative performance via community-wise competitions. Third, we provide dozens of Jupyter/Python demos organized into a curriculum and a documentation website to serve the rapidly growing community. The open-source codes for the data curation pipeline are available at https://github.com/AI4Finance-Foundation/FinRL-Meta

Ming Zhu, Xiao-Yang Liu, Xiaodong Wang

Public vehicle (PV) systems are promising transportation systems for future smart cities which provide dynamic ride-sharing services according to passengers' requests. PVs are driverless/self-driving electric vehicles which require frequent recharging from smart grids. For such systems, the challenge lies in both the efficient scheduling scheme to satisfy transportation demands with service guarantee and the cost-effective charging strategy under the real-time electricity pricing. In this paper, we study the joint transportation and charging scheduling for PV systems to balance the transportation and charging demands, ensuring the long-term operation. We adopt a cake cutting game model to capture the interactions among PV groups, the cloud and smart grids. The cloud announces strategies to coordinate the allocation of transportation and energy resources among PV groups. All the PV groups try to maximize their joint transportation and charging utilities. We propose an algorithm to obtain the unique normalized Nash equilibrium point for this problem. Simulations are performed to confirm the effects of our scheme under the real taxi and power grid data sets of New York City. Our results show that our scheme achieves almost the same transportation performance compared with a heuristic scheme, namely, transportation with greedy charging; however, the average energy price of the proposed scheme is 10.86% lower than the latter one.

Xiao-Yang Liu, Xiaodong Wang

The big data era is swamping areas including data analysis, machine/deep learning, signal processing, statistics, scientific computing, and cloud computing. The multidimensional feature and huge volume of big data put urgent requirements to the development of multilinear modeling tools and efficient algorithms. In this paper, we build a novel multilinear tensor space that supports useful algorithms such as SVD and QR, while generalizing the matrix space to fourth-order tensors was believed to be challenging. Specifically, given any multidimensional discrete transform, we show that fourth-order tensors are bilinear operators on a space of matrices. First, we take a transform-based approach to construct a new tensor space by defining a new multiplication operation and tensor products, and accordingly the analogous concepts: identity, inverse, transpose, linear combinations, and orthogonality. Secondly, we define the $\mathcal{L}$-SVD for fourth-order tensors and present an efficient algorithm, where the tensor case requires a stronger condition for unique decomposition than the matrix case. Thirdly, we define the tensor $\mathcal{L}$-QR decomposition and propose a Householder QR algorithm to avoid the catastrophic cancellation problem associated with the conventional Gram-Schmidt process. Finally, we validate our schemes on video compression and one-shot face recognition. For video compression, compared with the existing tSVD, the proposed $\mathcal{L}$-SVD achieves $3\sim 10$dB gains in RSE, while the running time is reduced by about $50\%$ and $87.5\%$, respectively. For one-shot face recognition, the recognition rate is increased by about $10\% \sim 20\%$.

Cuiping Li, Xiao-Yang Liu, Yue Sun

The problem of testing whether a signal lies within a given subspace, also named matched subspace detection, has been well studied when the signal is represented as a vector. However, the matched subspace detection methods based on vectors can not be applied to the situations that signals are naturally represented as multi-dimensional data arrays or tensors. Considering that tensor subspaces and orthogonal projections onto these subspaces are well defined in the recently proposed transform-based tensor model, which motivates us to investigate the problem of matched subspace detection in high dimensional case. In this paper, we propose an approach for tensor matched subspace detection based on the transform-based tensor model with tubal-sampling and elementwise-sampling, respectively. First, we construct estimators based on tubal-sampling and elementwise-sampling to estimate the energy of a signal outside a given subspace of a third-order tensor and then give the probability bounds of our estimators, which show that our estimators work effectively when the sample size is greater than a constant. Secondly, the detectors both for noiseless data and noisy data are given, and the corresponding detection performance analyses are also provided. Finally, based on discrete Fourier transform (DFT) and discrete cosine transform (DCT), the performance of our estimators and detectors are evaluated by several simulations, and simulation results verify the effectiveness of our approach.

Haoqiang Kang, Xiao-Yang Liu

The hallucination issue is recognized as a fundamental deficiency of large language models (LLMs), especially when applied to fields such as finance, education, and law. Despite the growing concerns, there has been a lack of empirical investigation. In this paper, we provide an empirical examination of LLMs' hallucination behaviors in financial tasks. First, we empirically investigate LLM model's ability of explaining financial concepts and terminologies. Second, we assess LLM models' capacity of querying historical stock prices. Third, to alleviate the hallucination issue, we evaluate the efficacy of four practical methods, including few-shot learning, Decoding by Contrasting Layers (DoLa), the Retrieval Augmentation Generation (RAG) method and the prompt-based tool learning method for a function to generate a query command. Finally, our major finding is that off-the-shelf LLMs experience serious hallucination behaviors in financial tasks. Therefore, there is an urgent need to call for research efforts in mitigating LLMs' hallucination.

Jifeng Li, Arnav Grover, Abraham Alpuerto et al.

The financial market is a mission-critical playground for AI agents due to its temporal dynamics and low signal-to-noise ratio. Building an effective algorithmic trading system may require a professional team to develop and test over the years. In this paper, we propose an orchestration framework for financial agents, which aims to democratize financial intelligence to the general public. We map each component of the traditional algorithmic trading system to agents, including planner, orchestrator, alpha agents, risk agents, portfolio agents, backtest agents, execution agents, audit agents, and memory agent. We present two in-house trading examples. For the stock trading task (hourly data from 04/2024 to 12/2024), our approach achieved a return of $20.42\%$, a Sharpe ratio of 2.63, and a maximum drawdown of $-3.59\%$, while the S&P 500 index yielded a return of $15.97\%$. For the BTC trading task (minute data from 27/07/2025 to 13/08/2025), our approach achieved a return of $8.39\%$, a Sharpe ratio of $0.38$, and a maximum drawdown of $-2.80\%$, whereas the BTC price increased by $3.80\%$. Our code is available on \href{https://github.com/Open-Finance-Lab/AgenticTrading}{GitHub}.

Xiao-Yang Liu, Hongyang Yang, Jiechao Gao et al.

Deep reinforcement learning (DRL) has been envisioned to have a competitive edge in quantitative finance. However, there is a steep development curve for quantitative traders to obtain an agent that automatically positions to win in the market, namely \textit{to decide where to trade, at what price} and \textit{what quantity}, due to the error-prone programming and arduous debugging. In this paper, we present the first open-source framework \textit{FinRL} as a full pipeline to help quantitative traders overcome the steep learning curve. FinRL is featured with simplicity, applicability and extensibility under the key principles, \textit{full-stack framework, customization, reproducibility} and \textit{hands-on tutoring}. Embodied as a three-layer architecture with modular structures, FinRL implements fine-tuned state-of-the-art DRL algorithms and common reward functions, while alleviating the debugging workloads. Thus, we help users pipeline the strategy design at a high turnover rate. At multiple levels of time granularity, FinRL simulates various markets as training environments using historical data and live trading APIs. Being highly extensible, FinRL reserves a set of user-import interfaces and incorporates trading constraints such as market friction, market liquidity and investor's risk-aversion. Moreover, serving as practitioners' stepping stones, typical trading tasks are provided as step-by-step tutorials, e.g., stock trading, portfolio allocation, cryptocurrency trading, etc.

Fanjie Kong, Xiao-yang Liu, Ricardo Henao

Tiny object classification problem exists in many machine learning applications like medical imaging or remote sensing, where the object of interest usually occupies a small region of the whole image. It is challenging to design an efficient machine learning model with respect to tiny object of interest. Current neural network structures are unable to deal with tiny object efficiently because they are mainly developed for images featured by large scale objects. However, in quantum physics, there is a great theoretical foundation guiding us to analyze the target function for image classification regarding to specific objects size ratio. In our work, we apply Tensor Networks to solve this arising tough machine learning problem. First, we summarize the previous work that connects quantum spin model to image classification and bring the theory into the scenario of tiny object classification. Second, we propose using 2D multi-scale entanglement renormalization ansatz (MERA) to classify tiny objects in image. In the end, our experimental results indicate that tensor network models are effective for tiny object classification problem and potentially will beat state-of-the-art. Our codes will be available online https://github.com/timqqt/MERA_Image_Classification.

Xiao-Yang Liu, Hongyang Yang, Qian Chen et al.

As deep reinforcement learning (DRL) has been recognized as an effective approach in quantitative finance, getting hands-on experiences is attractive to beginners. However, to train a practical DRL trading agent that decides where to trade, at what price, and what quantity involves error-prone and arduous development and debugging. In this paper, we introduce a DRL library FinRL that facilitates beginners to expose themselves to quantitative finance and to develop their own stock trading strategies. Along with easily-reproducible tutorials, FinRL library allows users to streamline their own developments and to compare with existing schemes easily. Within FinRL, virtual environments are configured with stock market datasets, trading agents are trained with neural networks, and extensive backtesting is analyzed via trading performance. Moreover, it incorporates important trading constraints such as transaction cost, market liquidity and the investor's degree of risk-aversion. FinRL is featured with completeness, hands-on tutorial and reproducibility that favors beginners: (i) at multiple levels of time granularity, FinRL simulates trading environments across various stock markets, including NASDAQ-100, DJIA, S&P 500, HSI, SSE 50, and CSI 300; (ii) organized in a layered architecture with modular structure, FinRL provides fine-tuned state-of-the-art DRL algorithms (DQN, DDPG, PPO, SAC, A2C, TD3, etc.), commonly-used reward functions and standard evaluation baselines to alleviate the debugging workloads and promote the reproducibility, and (iii) being highly extendable, FinRL reserves a complete set of user-import interfaces. Furthermore, we incorporated three application demonstrations, namely single stock trading, multiple stock trading, and portfolio allocation. The FinRL library will be available on Github at link https://github.com/AI4Finance-LLC/FinRL-Library.

Yunyao Mao, Xiaoyang Liu, Wengang Zhou et al.

Text-driven human motion generation, as one of the vital tasks in computer-aided content creation, has recently attracted increasing attention. While pioneering research has largely focused on improving numerical performance metrics on given datasets, practical applications reveal a common challenge: existing methods often overfit specific motion expressions in the training data, hindering their ability to generalize to novel descriptions like unseen combinations of motions. This limitation restricts their broader applicability. We argue that the aforementioned problem primarily arises from the scarcity of available motion-text pairs, given the many-to-many nature of text-driven motion generation. To tackle this problem, we formulate text-to-motion generation as a Markov decision process and present \textbf{InstructMotion}, which incorporate the trail and error paradigm in reinforcement learning for generalizable human motion generation. Leveraging contrastive pre-trained text and motion encoders, we delve into optimizing reward design to enable InstructMotion to operate effectively on both paired data, enhancing global semantic level text-motion alignment, and synthetic text-only data, facilitating better generalization to novel prompts without the need for ground-truth motion supervision. Extensive experiments on prevalent benchmarks and also our synthesized unpaired dataset demonstrate that the proposed InstructMotion achieves outstanding performance both quantitatively and qualitatively.

Dannong Wang, Daniel Kim, Bo Jin et al.

Finetuned large language models (LLMs) have shown remarkable performance in financial tasks, such as sentiment analysis and information retrieval. Due to privacy concerns, finetuning and deploying Financial LLMs (FinLLMs) locally are crucial for institutions. However, finetuning FinLLMs poses challenges including GPU memory constraints and long input sequences. In this paper, we employ quantized low-rank adaptation (QLoRA) to finetune FinLLMs, which leverage low-rank matrix decomposition and quantization techniques to significantly reduce computational requirements while maintaining high model performance. We also employ data and pipeline parallelism to enable local finetuning using cost-effective, widely accessible GPUs. Experiments on financial datasets demonstrate that our method achieves substantial improvements in accuracy, GPU memory usage, and time efficiency, underscoring the potential of lowrank methods for scalable and resource-efficient LLM finetuning.

Yan Wang, Yang Ren, Lingfei Qian et al.

Accurately understanding numbers from financial reports is fundamental to how markets, regulators, algorithms, and normal people read the economy and the world, yet even with XBRL (eXtensible Business Reporting Language) designed to tag every figure with standardized accounting concepts, mapping thousands of facts to over 10,000 U.S. GAAP concepts remains costly, inconsistent, and error-prone. Existing benchmarks define tagging as flat, single-step, extreme classification over small subsets of US-GAAP concepts, overlooking both the taxonomy's hierarchical semantics and the structured nature of real tagging, where each fact must be represented as a contextualized multi-field output. These simplifications prevent fair evaluation of large language models (LLMs) under realistic reporting conditions. To address these gaps, we introduce FinTagging, the first comprehensive benchmark for structure-aware and full-scope XBRL tagging, designed to evaluate LLMs' ability to extract and align financial facts through numerical reasoning and taxonomy alignment across text and tables. We define two subtasks: FinNI for numeric identification, which extracts numerical entities and their types from XBRL reports, and FinCL for concept linking, which maps each extracted entity to the corresponding concept in the full US-GAAP taxonomy. Together, these subtasks produce a structured representation of each financial fact. We evaluate diverse LLMs under zero-shot settings and analyze their performance across both subtasks and overall tagging accuracy. Results show that LLMs generalize well in numeric identification but struggle with fine-grained concept linking, revealing current limitations in structure-aware reasoning for accurate financial disclosure. All code and datasets are available on GitHub and Hugging Face.

Miao Yin, Siyu Liao, Xiao-Yang Liu et al.

Recurrent Neural Networks (RNNs) have been widely used in sequence analysis and modeling. However, when processing high-dimensional data, RNNs typically require very large model sizes, thereby bringing a series of deployment challenges. Although various prior works have been proposed to reduce the RNN model sizes, executing RNN models in resource-restricted environments is still a very challenging problem. In this paper, we propose to develop extremely compact RNN models with fully decomposed hierarchical Tucker (FDHT) structure. The HT decomposition does not only provide much higher storage cost reduction than the other tensor decomposition approaches but also brings better accuracy performance improvement for the compact RNN models. Meanwhile, unlike the existing tensor decomposition-based methods that can only decompose the input-to-hidden layer of RNNs, our proposed fully decomposition approach enables the comprehensive compression for the entire RNN models with maintaining very high accuracy. Our experimental results on several popular video recognition datasets show that our proposed fully decomposed hierarchical tucker-based LSTM (FDHT-LSTM) is extremely compact and highly efficient. To the best of our knowledge, FDHT-LSTM, for the first time, consistently achieves very high accuracy with only few thousand parameters (3,132 to 8,808) on different datasets. Compared with the state-of-the-art compressed RNN models, such as TT-LSTM, TR-LSTM and BT-LSTM, our FDHT-LSTM simultaneously enjoys both order-of-magnitude (3,985x to 10,711x) fewer parameters and significant accuracy improvement (0.6% to 12.7%).

Xiao-Yang Liu, Ming Zhu

Graph data completion is a fundamentally important issue as data generally has a graph structure, e.g., social networks, recommendation systems, and the Internet of Things. We consider a graph where each node has a data matrix, represented as a \textit{graph-tensor} by stacking the data matrices in the third dimension. In this paper, we propose a \textit{Convolutional Graph-Tensor Net} (\textit{Conv GT-Net}) for the graph data completion problem, which uses deep neural networks to learn the general transform of graph-tensors. The experimental results on the ego-Facebook data sets show that the proposed \textit{Conv GT-Net} achieves significant improvements on both completion accuracy (50\% higher) and completion speed (3.6x $\sim$ 8.1x faster) over the existing algorithms.

Daizong Liu, Xiaoye Qu, Xiao-Yang Liu et al.

Query-based moment localization is a new task that localizes the best matched segment in an untrimmed video according to a given sentence query. In this localization task, one should pay more attention to thoroughly mine visual and linguistic information. To this end, we propose a novel Cross- and Self-Modal Graph Attention Network (CSMGAN) that recasts this task as a process of iterative messages passing over a joint graph. Specifically, the joint graph consists of Cross-Modal interaction Graph (CMG) and Self-Modal relation Graph (SMG), where frames and words are represented as nodes, and the relations between cross- and self-modal node pairs are described by an attention mechanism. Through parametric message passing, CMG highlights relevant instances across video and sentence, and then SMG models the pairwise relation inside each modality for frame (word) correlating. With multiple layers of such a joint graph, our CSMGAN is able to effectively capture high-order interactions between two modalities, thus enabling a further precise localization. Besides, to better comprehend the contextual details in the query, we develop a hierarchical sentence encoder to enhance the query understanding. Extensive experiments on four public datasets demonstrate the effectiveness of our proposed model, and GCSMAN significantly outperforms the state-of-the-arts.

Wenhao Zhang, Wentian Bao, Xiao-Yang Liu et al.

Post-click conversion rate (CVR) estimation is a critical task in e-commerce recommender systems. This task is deemed quite challenging under the industrial setting with two major issues: 1) selection bias caused by user self-selection, and 2) data sparsity due to the rare click events. A successful conversion typically has the following sequential events: "exposure -> click -> conversion". Conventional CVR estimators are trained in the click space, but the inference is done in the entire exposure space. They fail to account for the causes of the missing data and treat them as missing at random. Hence, their estimations are highly likely to deviate from the real values by large. In addition, the data sparsity issue can also handicap many industrial CVR estimators which usually have large parameter spaces. In this paper, we propose two principled, efficient and highly effective CVR estimators for industrial CVR estimation, namely, Multi-IPW and Multi-DR. The proposed models approach the CVR estimation from a causal perspective and account for the causes of missing not at random. In addition, our methods are based on the multi-task learning framework and mitigate the data sparsity issue. Extensive experiments on industrial-level datasets show that our methods outperform the state-of-the-art CVR models.

Ming Zhu, Xiao-Yang Liu, Anwar Walid

Unmanned aerial vehicles (UAVs) are envisioned to complement the 5G communication infrastructure in future smart cities. Hot spots easily appear in road intersections, where effective communication among vehicles is challenging. UAVs may serve as relays with the advantages of low price, easy deployment, line-of-sight links, and flexible mobility. In this paper, we study a UAV-assisted vehicular network where the UAV jointly adjusts its transmission control (power and channel) and 3D flight to maximize the total throughput. First, we formulate a Markov decision process (MDP) problem by modeling the mobility of the UAV/vehicles and the state transitions. Secondly, we solve the target problem using a deep reinforcement learning method, namely, the deep deterministic policy gradient (DDPG), and propose three solutions with different control objectives. Deep reinforcement learning methods obtain the optimal policy through the interactions with the environment without knowing the environment variables. Considering that environment variables in our problem are unknown and unmeasurable, we choose a deep reinforcement learning method to solve it. Moreover, considering the energy consumption of 3D flight, we extend the proposed solutions to maximize the total throughput per unit energy. To encourage or discourage the UAV's mobility according to its prediction, the DDPG framework is modified, where the UAV adjusts its learning rate automatically. Thirdly, in a simplified model with small state space and action space, we verify the optimality of proposed algorithms. Comparing with two baseline schemes, we demonstrate the effectiveness of proposed algorithms in a realistic model.

Zihan Ding, Xiao-Yang Liu, Miao Yin et al.

Deep generative models have been successfully applied to many applications. However, existing works experience limitations when generating large images (the literature usually generates small images, e.g. 32 * 32 or 128 * 128). In this paper, we propose a novel scheme, called deep tensor adversarial generative nets (TGAN), that generates large high-quality images by exploring tensor structures. Essentially, the adversarial process of TGAN takes place in a tensor space. First, we impose tensor structures for concise image representation, which is superior in capturing the pixel proximity information and the spatial patterns of elementary objects in images, over the vectorization preprocess in existing works. Secondly, we propose TGAN that integrates deep convolutional generative adversarial networks and tensor super-resolution in a cascading manner, to generate high-quality images from random distributions. More specifically, we design a tensor super-resolution process that consists of tensor dictionary learning and tensor coefficients learning. Finally, on three datasets, the proposed TGAN generates images with more realistic textures, compared with state-of-the-art adversarial autoencoders. The size of the generated images is increased by over 8.5 times, namely 374 * 374 in PASCAL2.

Weijun Lu, Xiao-Yang Liu, Qingwei Wu et al.

We propose a novel multilinear dynamical system (MLDS) in a transform domain, named $\mathcal{L}$-MLDS, to model tensor time series. With transformations applied to a tensor data, the latent multidimensional correlations among the frontal slices are built, and thus resulting in the computational independence in the transform domain. This allows the exact separability of the multi-dimensional problem into multiple smaller LDS problems. To estimate the system parameters, we utilize the expectation-maximization (EM) algorithm to determine the parameters of each LDS. Further, $\mathcal{L}$-MLDSs significantly reduce the model parameters and allows parallel processing. Our general $\mathcal{L}$-MLDS model is implemented based on different transforms: discrete Fourier transform, discrete cosine transform and discrete wavelet transform. Due to the nonlinearity of these transformations, $\mathcal{L}$-MLDS is able to capture the nonlinear correlations within the data unlike the MLDS \cite{rogers2013multilinear} which assumes multi-way linear correlations. Using four real datasets, the proposed $\mathcal{L}$-MLDS is shown to achieve much higher prediction accuracy than the state-of-the-art MLDS and LDS with an equal number of parameters under different noise models. In particular, the relative errors are reduced by $50\% \sim 99\%$. Simultaneously, $\mathcal{L}$-MLDS achieves an exponential improvement in the model's training time than MLDS.

Ming Zhu, Xiao-Yang Liu, Xiaodong Wang

As efficient traffic-management platforms, public vehicle (PV) systems are envisioned to be a promising approach to solving traffic congestions and pollutions for future smart cities. PV systems provide online/dynamic peer-to-peer ride-sharing services with the goal of serving sufficient number of customers with minimum number of vehicles and lowest possible cost. A key component of the PV system is the online ride-sharing scheduling strategy. In this paper, we propose an efficient path planning strategy that focuses on a limited potential search area for each vehicle by filtering out the requests that violate passenger service quality level, so that the global search is reduced to local search. We analyze the performance of the proposed solution such as reduction ratio of computational complexity. Simulations based on the Manhattan taxi data set show that, the computing time is reduced by 22% compared with the exhaustive search method under the same service quality performance.

Fei Jiang, Xiao-Yang Liu, Hongtao Lu et al.

Sparse coding (SC) is an unsupervised learning scheme that has received an increasing amount of interests in recent years. However, conventional SC vectorizes the input images, which destructs the intrinsic spatial structures of the images. In this paper, we propose a novel graph regularized tensor sparse coding (GTSC) for image representation. GTSC preserves the local proximity of elementary structures in the image by adopting the newly proposed tubal-tensor representation. Simultaneously, it considers the intrinsic geometric properties by imposing graph regularization that has been successfully applied to uncover the geometric distribution for the image data. Moreover, the returned sparse representations by GTSC have better physical explanations as the key operation (i.e., circular convolution) in the tubal-tensor model preserves the shifting invariance property. Experimental results on image clustering demonstrate the effectiveness of the proposed scheme.

Cai Fu, Chenchen Peng, Xiao-Yang Liu

The search engine is tightly coupled with social networks and is primarily designed for users to acquire interested information. Specifically, the search engine assists the information dissemination for social networks, i.e., enabling users to access interested contents with keywords-searching and promoting the process of contents-transferring from the source users directly to potential interested users. Accompanying such processes, the social network evolves as new links emerge between users with common interests. However, there is no clear understanding of such a "chicken-and-egg" problem, namely, new links encourage more social interactions, and vice versa. In this paper, we aim to quantitatively characterize the social network evolution phenomenon driven by a search engine. First, we propose a search network model for social network evolution. Second, we adopt two performance metrics, namely, degree distribution and network diameter. Theoretically, we prove that the degree distribution follows an intensified power-law, and the network diameter shrinks. Third, we quantitatively show that the search engine accelerates the rumor propagation in social networks. Finally, based on four real-world data sets (i.e., CDBLP, Facebook, Weibo Tweets, P2P), we verify our theoretical findings. Furthermore, we find that the search engine dramatically increases the speed of rumor propagation.

Xiao-Yang Liu, Shuchin Aeron, Vaneet Aggarwal et al.

The low-tubal-rank tensor model has been recently proposed for real-world multidimensional data. In this paper, we study the low-tubal-rank tensor completion problem, i.e., to recover a third-order tensor by observing a subset of its elements selected uniformly at random. We propose a fast iterative algorithm, called {\em Tubal-Alt-Min}, that is inspired by a similar approach for low-rank matrix completion. The unknown low-tubal-rank tensor is represented as the product of two much smaller tensors with the low-tubal-rank property being automatically incorporated, and Tubal-Alt-Min alternates between estimating those two tensors using tensor least squares minimization. First, we note that tensor least squares minimization is different from its matrix counterpart and nontrivial as the circular convolution operator of the low-tubal-rank tensor model is intertwined with the sub-sampling operator. Second, the theoretical performance guarantee is challenging since Tubal-Alt-Min is iterative and nonconvex in nature. We prove that 1) Tubal-Alt-Min guarantees exponential convergence to the global optima, and 2) for an $n \times n \times k$ tensor with tubal-rank $r \ll n$, the required sampling complexity is $O(nr^2k \log^3 n)$ and the computational complexity is $O(n^2rk^2 \log^2 n)$. Third, on both synthetic data and real-world video data, evaluation results show that compared with tensor-nuclear norm minimization (TNN-ADMM), Tubal-Alt-Min improves the recovery error dramatically (by orders of magnitude). It is estimated that Tubal-Alt-Min converges at an exponential rate $10^{-0.4423 \text{Iter}}$ where $\text{Iter}$ denotes the number of iterations, which is much faster than TNN-ADMM's $10^{-0.0332 \text{Iter}}$, and the running time can be accelerated by more than $5$ times for a $200 \times 200 \times 20$ tensor.