With the booming development of multimodal data (e.g., image, text) on internet platforms, multimodal sequential recommendation methods continue to emerge. Most existing methods incorporate item modal features as auxiliary information, typically concatenating them to learn unified user representations. However, these methods directly use modal features for representation learning, neglecting the impact of inherent modal noise. We argue that internal-modal noise and cross-modal noise hinder the acquisition of more accurate user representations. To address this problem, we propose SGP4SR - Separated-modality Guided user Preference learning for multimodal Sequential Recommendation. Globally, the user preference modeling is carried out from a separated-modality perspective to alleviate cross-modal noise. Locally, for each individual modality, we use item relationship graphs and user interest centers, aggregated with ID embeddings, to replace direct modal features, thereby mitigating internal-modal noise. Finally, user representations from both separated-modality and multimodal perspectives participate in prediction independently. In experiments conducted on four real-world datasets, our method outperforms state-of-the-art approaches, achieving an average performance improvement of up to 8.84% over the best baseline. The comprehensive experiments further validate the superior noise tolerance and robustness of our method.

Recommendation model interpretation aims to reveal the relationships between inputs, model internal representations, and outputs to enhance the transparency, interpretability, and trustworthiness of recommendation systems. However, the inherent complexity and opacity of deep learning models pose challenges for model-level interpretation. Moreover, most existing methods for interpreting recommendation models are tailored to specific architectures or model types, limiting their generalizability across different types of recommenders. In this article, we propose RecSAE, a generalizable probing framework that interprets Rec ommendation models with S parse A uto E ncoders. The framework extracts interpretable latents from the internal representations of recommendation models and links them to semantic concepts for interpretations. It does not alter original models during interpretations and also enables targeted tuning to models. Experiments on three types of recommendation models (general, graph-based, sequential) with four widely used public datasets demonstrate the effectiveness and generalization of the RecSAE framework. The interpreted concepts are further validated by human experts, showing strong alignment with human perception. Overall, RecSAE serves as a novel step in both model-level interpretations to various types of recommendation models without affecting their functions and offering potential for targeted tuning of models. The code and data are available at https://github.com/Alice1998/RecSAE .

Despite their powerful text generation capabilities, large language models (LLMs) still struggle to effectively utilize external tools to solve complex tasks, a challenge known as tool learning. Existing methods primarily rely on supervised fine-tuning, treating tool learning as a text generation problem while overlooking the decision-making complexities inherent in multi-step contexts. In this work, we propose modeling tool learning as a dynamic decision-making process and introduce StepTool, a novel step-grained reinforcement learning framework that enhances LLMs' capabilities in multi-step tool use. StepTool comprises two key components: Step-grained Reward Shaping, which assigns rewards to each tool interaction based on its invocation success and contribution to task completion; and Step-grained Optimization, which applies policy gradient methods to optimize the model across multiple decision steps. Extensive experiments across diverse benchmarks show that StepTool consistently outperforms both SFT-based and RL-based baselines in terms of task Pass Rate and Recall of relevant tools. Furthermore, our analysis suggests that StepTool helps models discover new tool-use strategies rather than merely re-weighting prior knowledge. These results highlight the importance of fine-grained decision modeling in tool learning and establish StepTool as a general and robust solution for enhancing multi-step tool use in LLMs. Code and data are available at https://github.com/yuyq18/StepTool.

As evaluation designs of large language models may shape our trajectory toward artificial general intelligence, comprehensive and forward-looking assessment is essential.Existing benchmarks primarily assess static knowledge, while intelligence also entails the ability to rapidly learn from experience.To this end, we advocate for the evaluation of Test-time Learning, the capacity to improve performance in experience-based, reasoning-intensive tasks during test time.In this work, we propose semantic games as effective testbeds for evaluating test-time learning, due to their resistance to saturation and inherent demand for strategic reasoning.We introduce an objective evaluation framework that compares model performance under both limited and cumulative experience settings, and contains four forms of experience representation.To provide a comparative baseline, we recruit eight human participants to complete the same task.Results show that LLMs exhibit measurable test-time learning capabilities; however, their improvements are less stable under cumulative experience and progress more slowly than those observed in humans.These findings underscore the potential of LLMs as general-purpose learning machines, while also revealing a substantial intellectual gap between models and humans, irrespective of how well LLMs perform on static benchmarks.Code and data are available 1 ."Give a man a fish, and you feed him for a day; teach a man to fish, and you feed him for a lifetime."

Stickers, though small, are a highly condensed form of visual expression, ubiquitous across messaging platforms and embraced by diverse cultures, genders, and age groups. Despite their popularity, sticker retrieval remains an underexplored task due to the significant human effort and subjectivity involved in constructing high-quality sticker query datasets. Although large language models (LLMs) excel at general NLP tasks, they falter when confronted with the nuanced, intangible, and highly specific nature of sticker query generation. To address this challenge, we propose a threefold solution. First, we introduce Sticktionary, a gamified annotation framework designed to gather diverse, high-quality, and contextually resonant sticker queries. Second, we present StickerQueries, a multilingual sticker query dataset containing 1,115 English and 615 Chinese queries, annotated by over 60 contributors across 60+ hours. Lastly, Through extensive quantitative and qualitative evaluation, we demonstrate that our approach significantly enhances query generation quality, retrieval accuracy, and semantic understanding in the sticker domain. To support future research, we publicly release our multilingual dataset along with two fine-tuned query generation models.

Instant messaging with texts and stickers has become a widely adopted communication medium, enabling efficient expression of user semantics and emotions. With the increased use of stickers conveying information and feelings, sticker retrieval and recommendation has emerged as an important area of research. However, a major limitation in existing literature has been the lack of datasets capturing temporal and user-specific sticker interactions, which has hindered further progress in user modeling and sticker personalization. To address this, we introduce User-Sticker, a dataset that includes temporal and user anonymous ID across conversations. It is the largest publicly available sticker dataset to date, containing 22K unique users, 370K stickers, and 8.3M messages. The raw data was collected from a popular messaging platform from 67 conversations over 720 hours of crawling. All text and image data were carefully vetted for safety and privacy checks and modifications. Spanning 10 domains, the U-Sticker dataset captures rich temporal, multilingual, and cross-domain behaviors not previously available in other datasets. Extensive quantitative and qualitative experiments demonstrate U-Sticker's practical applications in user behavior modeling and personalized recommendation and highlight its potential to further research areas in personalized retrieval and conversational studies. U-Sticker dataset is publicly available.

The rapid growth of short videos has necessitated effective recommender systems to match users with content tailored to their evolving preferences. Current video recommendation models primarily treat each video as a whole, overlooking the dynamic nature of user preferences with specific video segments. In contrast, our research focuses on segment-level user interest modeling, which is crucial for understanding how users' preferences evolve during video browsing. To capture users' dynamic segment interests, we propose an innovative model that integrates a hybrid representation module, a multi-modal user-video encoder, and a segment interest decoder. Our model addresses the challenges of capturing dynamic interest patterns, missing segment-level labels, and fusing different modalities, achieving precise segment-level interest prediction.

With the development of multi-modal modeling techniques, recent sequential recommender systems enhance transferability by incorporating cross-domain universal multi-modal data, e.g., text and image. Existing methods typically adopt pairwise alignment to alleviate the gap between modalities. However, this alignment paradigm has limitations on explainability, consistency, and expansibility, resulting in suboptimal performance. This paper proposes a novel Explainable multi-modality Alignment method for transferable Rec ommender systems, i.e., EARec. Specifically, we design a two-stage framework to achieve explainable modality alignment in the source domain and recommendation based on aligned modality representations in the target domain. In the first stage, we adopt a generative task to align various modalities in parallel to a shared anchor with explainable meaning. All modalities share the same anchor to ensure consistent direction. Additionally, we treat behavior as an independent modality to integrate task-specific information into the alignment framework. In the second stage, we compose multiple item modality representation models trained in the first stage to obtain a unified model capable of understanding various modalities simultaneously, thereby providing high-quality item modality representations for recommendations in the target domain. Benefiting from the approach of parallel modality alignment followed by model composition, the framework shows flexibility in expanding new modalities. Experimental results on multiple public datasets demonstrate the superiority of EARec over baselines, and further analyses indicate the explainability and expansibility of the proposed alignment method.

Instant Messaging is a popular mean for daily communication, allowing users to send text and stickers.As the saying goes, "a picture is worth a thousand words", so developing an effective sticker retrieval technique is crucial for enhancing user experience.However, existing sticker retrieval methods rely on labeled data to interpret stickers, and general-purpose Vision-Language Models (VLMs) often struggle to capture the unique semantics of stickers.Additionally, relevance-based sticker retrieval methods lack personalization, creating a gap between diverse user expectations and retrieval results.To address these, we propose the Personalized Sticker Retrieval with Vision-Language Model framework, namely PerSRV, structured into offline calculations and online processing modules.The online retrieval part follows the paradigm of relevant recall and personalized ranking, supported by the offline precalculation parts, which are sticker semantic understanding, utility evaluation and personalization modules.Firstly, for sticker-level semantic understanding, we supervised fine-tuned LLaVA-1.5-7B to generate human-like sticker semantics, complemented by textual content extracted from figures and historical interaction queries.Secondly, we investigate three crowd-sourcing metrics for sticker utility evaluation.Thirdly, we cluster style centroids based on users' historical interactions to achieve personal preference modeling.Finally, we evaluate our proposed PerSRV method on a public sticker retrieval dataset from WeChat 1 , containing 543,098 candidates and 12,568 interactions.Experimental results show that PerSRV significantly outperforms existing methods in multi-modal sticker retrieval.Additionally, our supervised fine-tuned VLM delivers notable improvements in sticker semantic understandings.The code 2 * Both authors contributed equally to this research.

The multimodal recommendation has gradually become the infrastructure of online media platforms, enabling them to provide personalized service to users through a joint modeling of user historical behaviors (e.g., purchases, clicks) and item various modalities (e.g., visual and textual). The majority of existing studies typically focus on utilizing modal features or modal-related graph structure to learn user local interests. Nevertheless, these approaches encounter two limitations: (1) Shared updates of user ID embeddings result in the consequential coupling between collaboration and multimodal signals; (2) Lack of exploration into robust global user interests to alleviate the sparse interaction problems faced by local interest modeling. To address these issues, we propose a novel Local and Global Graph Learning-guided Multimodal Recommender (LGMRec), which jointly models local and global user interests. Specifically, we present a local graph embedding module to independently learn collaborative-related and modality-related embeddings of users and items with local topological relations. Moreover, a global hypergraph embedding module is designed to capture global user and item embeddings by modeling insightful global dependency relations. The global embeddings acquired within the hypergraph embedding space can then be combined with two decoupled local embeddings to improve the accuracy and robustness of recommendations. Extensive experiments conducted on three benchmark datasets demonstrate the superiority of our LGMRec over various state-of-the-art recommendation baselines, showcasing its effectiveness in modeling both local and global user interests.

Learning precise representations of users and items to fit observed interaction data is the fundamental task of collaborative filtering. Existing studies usually infer entangled representations to fit such interaction data, neglecting to model the diverse matching relationships between users and items behind their interactions, leading to limited performance and weak interpretability. To address this problem, we propose a Dual Disentangled Variational AutoEncoder (DualVAE) for collaborative recommendation, which combines disentangled representation learning with variational inference to facilitate the generation of implicit interaction data. Specifically, we first implement the disentangling concept by unifying an attention-aware dual disentanglement and disentangled variational autoencoder to infer the disentangled latent representations of users and items. Further, to encourage the correspondence and independence of disentangled representations of users and items, we design a neighborhood-enhanced representation constraint with a customized contrastive mechanism to improve the representation quality. Extensive experiments on three real-world benchmarks show that our proposed model significantly outperforms several recent state-of-the-art baselines. Further empirical experimental results also illustrate the interpretability of the disentangled representations learned by DualVAE.

Session-based recommendation (SBR) aims to predict the anonymous user's next-click items by modeling the short-term sequence pattern. As most existing SBR models generally generate item representations based only on information propagation over the short sequence while ignoring additional valuable knowledge, their expressive abilities are somewhat limited by data sparsity caused by short sequence. Though there have been some attempts on utilizing items' attributes, they basically embed attributes into items directly, ignoring the fact that 1) there is no contextual relationship among attributes; and 2) users have varying levels of attention to different attributes, which still leads to unsatisfactory performance. To tackle the issues, we propose a novel Attribute-enhanced Dual Channel Representation Learning (ADRL) model for SBR, in which we independently model session representations in attribute-related pattern and sequence-related pattern. Specifically, we learn session representations with sequence patterns from the session graph, and we further design an frequency-driven attribute aggregator to generate the attribute-related session representations within a session. The proposed attribute aggregator is plug-and-play, as it can be coupled with most existing SBR models. Extensive experiments on three real-world public datasets demonstrate the superiority of the proposed ADRL over several state-of-the-art baselines, as well as the effectiveness and efficiency of our attribute aggregator module.

Next-basket recommendation (NBR) is a type of recommendation task that focuses on mining user interests based on the sequential basket records in which users purchase multiple items at a time. Limited by the sparsity brought by short-term user interaction behavior, existing NBR methods typically fail to mine fine-grained and complete representation of user interests, resulting in unsatisfactory recommendation performance. To address this issue, we propose a novel Multi-aspect Interest Neighbor-augmented Network (MINN) to capture fine-grained and complete representation of user interest for the next basket prediction. Specifically, we first design a multi-aspect interest encoder to learn representations of users and items in the multi-aspect interest space. Semantic neighbors are then selected for user to enhance the user’s representation of interest through a semantic neighbor augmentation mechanism. Finally, the user’s representation of interest and his/her repeated purchase behavior are jointly considered to implement the next basket prediction. Extensive experimental studies on two benchmark datasets demonstrate that MINN outperforms several representative NBR methods and achieves new state-of-the-art results.

This paper describes our submission to the ICASSP 2023 Signal Processing Grand Challenge (SPGC), which focuses on multilingual Alzheimer’s disease (AD) recognition through spontaneous speech. Our approaches include using a variety of acoustic features and silence-related information for AD detection and mini-mental state examination (MMSE) score prediction, and fine-tuning wav2vec2.0 models on speech in various frequency bands for AD detection. Our overall results on the test data outperform the baseline provided by the organizers, achieving 73.9% accuracy in AD detection by fine-tuning our bilingual wav2vec2.0 pre-trained model on the 0-1000Hz frequency band speech, and 4.610 RMSE (r = 0.565) in MMSE prediction through the fusion of eGeMAPS and silence features.

Cross-domain collaborative filtering (CDCF) is an effective method to alleviate the data sparsity problem by transferring knowledge from a source domain to assist the learning of a target domain. However, most of the existing CDCF approaches require that the two domains have at least one overlapping side (either on user or item) and the raw data can be fully shared across domains, which is difficult to be satisfied in reality due to corporate barriers and the risk of user privacy leakage. Although there are some attempts on applying CDCF to the scenario without overlapping data by transferring cluster-level rating patterns, these methods fail to mine the complex connections between the two domains, which makes their performance still not satisfactory. To address these problems, we propose a novel deep Interaction Distribution Transfer (IDT) framework, which extracts and transfers knowledge from the feature distribution formed by the whole dataset rather than specific data. In this way, the knowledge is embedded into high-order features for transfer, which can effectively avoid privacy leakage during the data sharing process. Moreover, as a flexible framework, IDT obtains powerful feature extraction ability from the base model, which guarantees its superior performance. Extensive experiments on three benchmark datasets are conducted and the results verify the effectiveness of the proposed framework.

Session-based recommendation (SBR) has received increasing attention to predict the next item via extracting and integrating both global and local item-item relationships. However, there still exist some deficiencies in current works when capturing these two kinds of relationships. For global item-item relationships, the global graph constructed by most SBR is a pseudo-global graph, which may cause redundant mining of sequence relationships. For local item-item relationships, conventional SBR only mines the sequence patterns while ignoring the feature patterns, which may introduce noise when learning users' interests. To address these problems, we propose a novel Knowledge-enhanced Multi-View Graph Neural Network (KMVG) by constructing three views, namely knowledge view, session view, and pairwise view. Specifically, benefiting from the rich semantic information in the knowledge graph (KG), we build a genuine global graph that is sequence-independent based on KG to mine the global item-item relationships in the knowledge view. Then, a session view is utilized to capture the contextual transitions among items as the sequence patterns of local item-item relationships, and a pairwise view is used to explore the feature commonality within a session as the feature patterns of the local item-item relationships. Extensive experiments on three real-world public datasets demonstrate the superiority of KMVG, showing that it outperforms the state-of-the-art baselines. Further analysis also reveals the effectiveness of KMVG in exploiting the item-item relationships under multiple views.

Next-basket recommendation (NBR) is a type of recommendation that aims to recommend a set of items to users according to their historical basket sequences. Existing NBR methods suffer from two limitations: (1) overlooking low-level item correlations, which results in coarse-grained item representation; and (2) failing to consider spurious interests in repeated behaviors, leading to suboptimal user interest learning. To address these limitations, we propose a novel solution named Multi-view Multi-aspect Neural Recommendation (MMNR) for NBR, which first normalizes the interactions from both the user-side and item-side, respectively, aiming to remove the spurious interests, and utilizes them as weights for items from different views to construct differentiated representations for each interaction item, enabling comprehensive user interest learning. Then, to capture low-level item correlations, MMNR models different aspects of items to obtain disentangled representations of items, thereby fully capturing multiple user interests. Extensive experiments on real-world datasets demonstrate the effectiveness of MMNR, showing that it consistently outperforms several state-of-the-art NBR methods.

Many studies have utilized speech data to automatically detect Alzheimer's Disease (AD). However, most of them simply take the speech data from a subject participating in a task (e.g., picture description) as a whole audio sequence, and lack considerations on separate utterances that may be related to different topics in the picture and are of different importance for discriminating AD patients from healthy controls. To this end, this paper proposes an AD detection method with topic modeling for utterances composed of audio words. First, an audio word discovery algorithm using Byte Pair Encoding (BPE) is designed to tokenize speech data without relying on text transcriptions. Then, a topic model is built that assigns each utterance a topic label in an unsupervised way. Finally, an Audio-Word HuBERT (AW-HuBERT) model integrating utterance-level topic labels is constructed for AD detection. This model is pretrained by training the existing HuBERT model with audio word sequences. The final decision for a recording is made by the weighted sum of utterance-level classification possibilities, and topic-dependent Area-Under-Curve (AUC) values are used as the weights. Experimental results on the DementiaBank dataset show that the proposed method achieves a better AD detection accuracy than state-of-the-art methods.

Due to the success of graph convolutional networks (GCNs) in effectively extracting features in non-Euclidean spaces, GCNs has become the rising star in implicit collaborative filtering. Existing works, while encouraging, typically adopt simple aggregation operation on the user-item bipartite graph to model user and item representations, but neglect to mine the sufficient dependencies between nodes, e.g., the relationships between users/items and their neighbors (or congeners), resulting in inadequate graph representation learning. To address these problems, we propose a novel Multi-Dependency Graph Collaborative Filtering (MDGCF) model, which mines the neighborhood- and homogeneous-level dependencies to enhance the representation power of graph-based CF models. Specifically, for neighborhood-level dependencies, we explicitly consider both popularity score and preference correlation by designing a joint neighborhood-level dependency weight, based on which we construct a neighborhood-level dependencies graph to capture higher-order interaction features. Besides, by adaptively mining the homogeneous-level dependencies among users and items, we construct two homogeneous graphs, based on which we further aggregate features from homogeneous users and items to supplement their representations, respectively. Extensive experiments on three real-world benchmark datasets demonstrate the effectiveness of the proposed MDGCF. Further experiments reveal that our model can capture rich dependencies between nodes for explaining user behaviors.

Neural network models have achieved state-of-the-art performance on grapheme-to-phoneme (G2P) conversion. However, their performance relies on large-scale pronunciation dictionaries, which may not be available for a lot of languages. Inspired by the success of the pre-trained language model BERT, this paper proposes a pre-trained grapheme model called grapheme BERT (GBERT), which is built by self-supervised training on a large, language-specific word list with only grapheme information. Furthermore, two approaches are developed to incorporate GBERT into the state-of-the-art Transformer-based G2P model, i.e., fine-tuning GBERT or fusing GBERT into the Transformer model by attention. Experimental results on the Dutch, Serbo-Croatian, Bulgarian and Korean datasets of the SIGMORPHON 2021 G2P task confirm the effectiveness of our GBERT-based G2P models under both medium-resource and low-resource data conditions.

This paper proposes a method of detecting dementia from the simultaneous speech and eye-tracking recordings of subjects in a picture description task. First, automatic speech recognition (ASR) and regional picture recognition (RPR) models are built to extract content-related bottleneck (BN) features for both speech and eye-tracking inputs. Then, a neural network is designed to fuse these two modals for discriminating dementia patients from healthy controls. The network contains a cross-modal Transformer encoder for bimodal interaction, and a self-attention Transformer encoder for final classification. Experimental results demonstrate that the detection accuracy of the proposed method is 84.26%, which outperforms baseline methods and ablated models using single speech or eye-tracking input.

Learning disentangled representations that reflect user preference based on user behavior (implicit feedback, such as click and purchase) and content information (e.g., plot description, poster) has become a hot research topic in modern recommender systems. However, most existing methods considering content information are not well-designed to disentangle user preference features due to neglecting the diversity of user preference on different semantic topics of items, resulting in sub-optimal performance and low interpretability. To address this problem, we propose a novelTopic-aware Disentangled Variational AutoEncoder (TopicVAE) to learn disentangled representations for enhanced recommendation. Specifically, we first utilize an attention-based topic extraction to extract the topic-level item representations and topic-item probability distribution from item content, and then introduce variational autoencoder to infer topic-level disentangled user representations. To guide the learning of topic-level disentanglement, we present a topic-guided self-supervised contrastive loss to promote the otherness of different topics by introducing a neighborhood-based user representation as guidance. Besides, a heuristic regularization is designed to force each dimension of the disentangled representations to independently reflect a fine-grained factor of a specific topic (e.g., red or blue for color) for feature-level disentanglement. Extensive experimental studies on three public datasets show that TopicVAE significantly outperforms several state-of-the-art baselines. Further empirical experiments also illustrate the interpretability of disentangled representations learned by TopicVAE.

This paper presents a method of detecting Alzheimer’s disease (AD) from the spontaneous speech of subjects in a picture description task using neural networks. This method does not rely on the manual transcriptions and annotations of a subject’s speech, but utilizes the bottleneck features extracted from audio using an ASR model. The neural network contains convolutional neural network (CNN) layers for local context modeling, bidirectional long shortterm memory (BiLSTM) layers for global context modeling and an attention pooling layer for classification. Furthermore, a masking- based data augmentation method is designed to deal with the data scarcity problem. Experiments on the DementiaBank dataset show that the detection accuracy of our proposed method is 82.59%, which is better than the baseline method based on manually-designed acoustic features and support vector machines (SVM), and achieves the state-of-the-art performance of detecting AD using only audio data on this dataset.

Data scarcity is always a constraint on analyzing speech transcriptions for automatic Alzheimer's disease (AD) detection, especially when the subjects are non-English speakers.To deal with this issue, this paper first proposes a contrastive learning method to obtain effective representations for text classification based on monolingual embeddings of BERT.Furthermore, a cross-lingual data augmentation method is designed by building autoencoders to learn the text representations shared by both languages.Experiments on a Mandarin AD corpus show that the contrastive learning method can achieve better detection accuracy than conventional CNN-based and BERTbased methods.Our cross-lingual data augmentation method also outperforms other compared methods when using another English AD corpus for augmentation.Finally, a best detection accuracy of 81.6% is obtained by our proposed methods on the Mandarin AD corpus.

Due to its nature of learning from dynamic interactions and planning for long-run performance, reinforcement learning (RL) recently has received much attention in interactive recommender systems (IRSs). IRSs usually face the large discrete action space problem, which makes most of the existing RL-based recommendation methods inefficient. Moreover, data sparsity is another challenging problem that most IRSs are confronted with. While the textual information like reviews and descriptions is less sensitive to sparsity, existing RL-based recommendation methods either neglect or are not suitable for incorporating textual information. To address these two problems, in this paper, we propose TDDPG-Rec, a Text-based Deep Deterministic Policy Gradient framework for interactive recommendation. Specifically, we leverage textual information to map items and users into a feature space, which greatly alleviates the sparsity problem. Moreover, we design an effective method to construct an action candidate set. By the policy vector dynamically learned from TDDPG-Rec that expresses the user's preference, we can select actions from the candidate set effectively. Through extensive experiments on three public datasets, we demonstrate that TDDPG-Rec achieves state-of-the-art performance over several baselines in a time-efficient manner.

Traditional schemes achieving load balancing in asymmetric topology, which need to maintain global or local congestion information, turn out to be complicated to implement. One recent research has verified that flowlet switching is more simple and efficient to achieve adaptive load balancing in asymmetric topology. Nevertheless, one tricky problem lies in determining the flowlet timeout value, δ. Setting it too small would risk reordering issue while setting it too large would reduce flowlet opportunities. In this paper, by formulating the timeout setting problem with a stationary distribution of Markov chain, we give a theoretical reference for setting an appropriate timeout value in flowlet switching based load balancing scheme. Then, we implement a flowlet switching based load balancing scheme, called EasyLB, by extending OpenFlow protocol. Experiment results show that, by setting timeout value following the preceding theoretical reference, EasyLB is adaptive to asymmetric topology and achieves fast convergence of load balancing after link failures.