IKFST: IOO and KOO Algorithms for Accelerated and Precise WFST-based End-to-End Automatic Speech Recognition

End-to-end automatic speech recognition has become the dominant paradigm in both academia and industry. To enhance recognition performance, the Weighted Finite-State Transducer (WFST) is widely adopted to integrate acoustic and language models through static graph composition, providing robust decoding and effective error correction. However, WFST decoding relies on a frame-by-frame autoregressive search over CTC posterior probabilities, which severely limits inference efficiency. Motivated by establishing a more principled compatibility between WFST decoding and CTC modeling, we systematically study the two fundamental components of CTC outputs, namely blank and non-blank frames, and identify a key insight: blank frames primarily encode positional information, while non-blank frames carry semantic content. Building on this observation, we introduce Keep-Only-One and Insert-Only-One, two decoding algorithms that explicitly exploit the structural roles of blank and non-blank frames to achieve significantly faster WFST-based inference without compromising recognition accuracy. Experiments on large-scale in-house, AISHELL-1, and LibriSpeech datasets demonstrate state-of-the-art recognition accuracy with substantially reduced decoding latency, enabling truly efficient and high-performance WFST decoding in modern speech recognition systems.

CoDA: A Context-Decoupled Hierarchical Agent with Reinforcement Learning

Large Language Model (LLM) agents trained with reinforcement learning (RL) show great promise for solving complex, multi-step tasks. However, their performance is often crippled by "Context Explosion", where the accumulation of long text outputs overwhelms the model's context window and leads to reasoning failures. To address this, we introduce CoDA, a Context-Decoupled hierarchical Agent, a simple but effective reinforcement learning framework that decouples high-level planning from low-level execution. It employs a single, shared LLM backbone that learns to operate in two distinct, contextually isolated roles: a high-level Planner that decomposes tasks within a concise strategic context, and a low-level Executor that handles tool interactions in an ephemeral, isolated workspace. We train this unified agent end-to-end using PECO (Planner-Executor Co-Optimization), a reinforcement learning methodology that applies a trajectory-level reward to jointly optimize both roles, fostering seamless collaboration through context-dependent policy updates. Extensive experiments demonstrate that CoDA achieves significant performance improvements over state-of-the-art baselines on complex multi-hop question-answering benchmarks, and it exhibits strong robustness in long-context scenarios, maintaining stable performance while all other baselines suffer severe degradation, thus further validating the effectiveness of our hierarchical design in mitigating context overload.

Enhancing Agentic RL with Progressive Reward Shaping and Value-based Sampling Policy Optimization

Large Language Models (LLMs) empowered with Tool-Integrated Reasoning (TIR) can iteratively plan, call external tools, and integrate returned information to solve complex, long-horizon reasoning tasks. Agentic Reinforcement Learning (Agentic RL) optimizes such models over full tool-interaction trajectories, but two key challenges hinder effectiveness: (1) Sparse, non-instructive rewards, such as binary 0-1 verifiable signals, provide limited guidance for intermediate steps and slow convergence; (2) Gradient degradation in Group Relative Policy Optimization (GRPO), where identical rewards within a rollout group yield zero advantage, reducing sample efficiency and destabilizing training. To address these challenges, we propose two complementary techniques: Progressive Reward Shaping (PRS) and Value-based Sampling Policy Optimization (VSPO). PRS is a curriculum-inspired reward design that introduces dense, stage-wise feedback - encouraging models to first master parseable and properly formatted tool calls, then optimize for factual correctness and answer quality. We instantiate PRS for short-form QA (with a length-aware BLEU to fairly score concise answers) and long-form QA (with LLM-as-a-Judge scoring to prevent reward hacking). VSPO is an enhanced GRPO variant that replaces low-value samples with prompts selected by a task-value metric balancing difficulty and uncertainty, and applies value-smoothing clipping to stabilize gradient updates. Experiments on multiple short-form and long-form QA benchmarks show that PRS consistently outperforms traditional binary rewards, and VSPO achieves superior stability, faster convergence, and higher final performance compared to PPO, GRPO, CISPO, and SFT-only baselines. Together, PRS and VSPO yield LLM-based TIR agents that generalize better across domains.

NAM: A Normalization Attention Model for Personalized Product Search In Fliggy

Personalized product search provides significant benefits to e-commerce platforms by extracting more accurate user preferences from historical behaviors. Previous studies largely focused on the user factors when personalizing the search query, while ignoring the item perspective, which leads to the following two challenges that we summarize in this paper: First, previous approaches relying only on co-occurrence frequency tend to overestimate the conversion rates for popular items and underestimate those for long-tail items, resulting in inaccurate item similarities; Second, user purchasing propensity is highly heterogeneous according to the popularity of the target item: it is less correlated with the user's historical behavior for a popular item and more correlated for a long-tail item. To address these challenges, in this paper we propose NAM, a Normalization Attention Model, which optimizes ''when to personalize'' by utilizing Inverse Item Frequency (IIF) and employing a gating mechanism, as well as optimizes ''how to personalize'' by normalizing the attention mechanism from a global perspective. Through comprehensive experiments, we demonstrate that our proposed NAM model significantly outperforms state-of-the-art baseline models. Furthermore, we conducted an online A/B test at Fliggy, and obtained a significant improvement of 0.8% over the latest production system in conversion rate.

CRAFT: Time Series Forecasting with Cross-Future Behavior Awareness

The past decades witness the significant advancements in time series forecasting (TSF) across various real-world domains, including e-commerce and disease spread prediction. However, TSF is usually constrained by the uncertainty dilemma of predicting future data with limited past observations. To settle this question, we explore the use of Cross-Future Behavior (CFB) in TSF, which occurs before the current time but takes effect in the future. We leverage CFB features and propose the CRoss-Future Behavior Awareness based Time Series Forecasting method (CRAFT). The core idea of CRAFT is to utilize the trend of cross-future behavior to mine the trend of time series data to be predicted. Specifically, to settle the sparse and partial flaws of cross-future behavior, CRAFT employs the Koopman Predictor Module to extract the key trend and the Internal Trend Mining Module to supplement the unknown area of the cross-future behavior matrix. Then, we introduce the External Trend Guide Module with a hierarchical structure to acquire more representative trends from higher levels. Finally, we apply the demand-constrained loss to calibrate the distribution deviation of prediction results. We conduct experiments on real-world dataset. Experiments on both offline large-scale dataset and online A/B test demonstrate the effectiveness of CRAFT. Our dataset and code is available at https://github.com/CRAFTinTSF/CRAFT.