Best Practices for Distilling Large Language Models into BERT for Web Search Ranking

Recent studies have highlighted the significant potential of Large Language Models (LLMs) as zero-shot relevance rankers. These methods predominantly utilize prompt learning to assess the relevance between queries and documents by generating a ranked list of potential documents. Despite their promise, the substantial costs associated with LLMs pose a significant challenge for their direct implementation in commercial search systems. To overcome this barrier and fully exploit the capabilities of LLMs for text ranking, we explore techniques to transfer the ranking expertise of LLMs to a more compact model similar to BERT, using a ranking loss to enable the deployment of less resource-intensive models. Specifically, we enhance the training of LLMs through Continued Pre-Training, taking the query as input and the clicked title and summary as output. We then proceed with supervised fine-tuning of the LLM using a rank loss, assigning the final token as a representative of the entire sentence. Given the inherent characteristics of autoregressive language models, only the final token </s> can encapsulate all preceding tokens. Additionally, we introduce a hybrid point-wise and margin MSE loss to transfer the ranking knowledge from LLMs to smaller models like BERT. This method creates a viable solution for environments with strict resource constraints. Both offline and online evaluations have confirmed the efficacy of our approach, and our model has been successfully integrated into a commercial web search engine as of February 2024.

CaPo: Cooperative Plan Optimization for Efficient Embodied Multi-Agent Cooperation

In this work, we address the cooperation problem among large language model (LLM) based embodied agents, where agents must cooperate to achieve a common goal. Previous methods often execute actions extemporaneously and incoherently, without long-term strategic and cooperative planning, leading to redundant steps, failures, and even serious repercussions in complex tasks like search-and-rescue missions where discussion and cooperative plan are crucial. To solve this issue, we propose Cooperative Plan Optimization (CaPo) to enhance the cooperation efficiency of LLM-based embodied agents. Inspired by human cooperation schemes, CaPo improves cooperation efficiency with two phases: 1) meta-plan generation, and 2) progress-adaptive meta-plan and execution. In the first phase, all agents analyze the task, discuss, and cooperatively create a meta-plan that decomposes the task into subtasks with detailed steps, ensuring a long-term strategic and coherent plan for efficient coordination. In the second phase, agents execute tasks according to the meta-plan and dynamically adjust it based on their latest progress (e.g., discovering a target object) through multi-turn discussions. This progress-based adaptation eliminates redundant actions, improving the overall cooperation efficiency of agents. Experimental results on the ThreeDworld Multi-Agent Transport and Communicative Watch-And-Help tasks demonstrate that CaPo achieves much higher task completion rate and efficiency compared with state-of-the-arts.

Hunyuan-Large: An Open-Source MoE Model with 52 Billion Activated Parameters by Tencent

In this paper, we introduce Hunyuan-Large, which is currently the largest open-source Transformer-based mixture of experts model, with a total of 389 billion parameters and 52 billion activation parameters, capable of handling up to 256K tokens. We conduct a thorough evaluation of Hunyuan-Large's superior performance across various benchmarks including language understanding and generation, logical reasoning, mathematical problem-solving, coding, long-context, and aggregated tasks, where it outperforms LLama3.1-70B and exhibits comparable performance when compared to the significantly larger LLama3.1-405B model. Key practice of Hunyuan-Large include large-scale synthetic data that is orders larger than in previous literature, a mixed expert routing strategy, a key-value cache compression technique, and an expert-specific learning rate strategy. Additionally, we also investigate the scaling laws and learning rate schedule of mixture of experts models, providing valuable insights and guidances for future model development and optimization. The code and checkpoints of Hunyuan-Large are released to facilitate future innovations and applications. Codes: https://github.com/Tencent/Hunyuan-Large Models: https://huggingface.co/tencent/Tencent-Hunyuan-Large

The NPU-HWC System for the ISCSLP 2024 Inspirational and Convincing Audio Generation Challenge

This paper presents the NPU-HWC system submitted to the ISCSLP 2024 Inspirational and Convincing Audio Generation Challenge 2024 (ICAGC). Our system consists of two modules: a speech generator for Track 1 and a background audio generator for Track 2. In Track 1, we employ Single-Codec to tokenize the speech into discrete tokens and use a language-model-based approach to achieve zero-shot speaking style cloning. The Single-Codec effectively decouples timbre and speaking style at the token level, reducing the acoustic modeling burden on the autoregressive language model. Additionally, we use DSPGAN to upsample 16 kHz mel-spectrograms to high-fidelity 48 kHz waveforms. In Track 2, we propose a background audio generator based on large language models (LLMs). This system produces scene-appropriate accompaniment descriptions, synthesizes background audio with Tango 2, and integrates it with the speech generated by our Track 1 system. Our submission achieves the second place and the first place in Track 1 and Track 2 respectively.

HASN: Hybrid Attention Separable Network for Efficient Image Super-resolution

Recently, lightweight methods for single image super-resolution (SISR) have gained significant popularity and achieved impressive performance due to limited hardware resources. These methods demonstrate that adopting residual feature distillation is an effective way to enhance performance. However, we find that using residual connections after each block increases the model's storage and computational cost. Therefore, to simplify the network structure and learn higher-level features and relationships between features, we use depthwise separable convolutions, fully connected layers, and activation functions as the basic feature extraction modules. This significantly reduces computational load and the number of parameters while maintaining strong feature extraction capabilities. To further enhance model performance, we propose the Hybrid Attention Separable Block (HASB), which combines channel attention and spatial attention, thus making use of their complementary advantages. Additionally, we use depthwise separable convolutions instead of standard convolutions, significantly reducing the computational load and the number of parameters while maintaining strong feature extraction capabilities. During the training phase, we also adopt a warm-start retraining strategy to exploit the potential of the model further. Extensive experiments demonstrate the effectiveness of our approach. Our method achieves a smaller model size and reduced computational complexity without compromising performance. Code can be available at https://github.com/nathan66666/HASN.git

Determine-Then-Ensemble: Necessity of Top-k Union for Large Language Model Ensembling

Large language models (LLMs) exhibit varying strengths and weaknesses across different tasks, prompting recent studies to explore the benefits of ensembling models to leverage their complementary advantages. However, existing LLM ensembling methods often overlook model compatibility and struggle with inefficient alignment of probabilities across the entire vocabulary. In this study, we empirically investigate the factors influencing ensemble performance, identifying model performance, vocabulary size, and response style as key determinants, revealing that compatibility among models is essential for effective ensembling. This analysis leads to the development of a simple yet effective model selection strategy that identifies compatible models. Additionally, we introduce the \textsc{Uni}on \textsc{T}op-$k$ \textsc{E}nsembling (\textsc{UniTE}), a novel approach that efficiently combines models by focusing on the union of the top-k tokens from each model, thereby avoiding the need for full vocabulary alignment and reducing computational overhead. Extensive evaluations across multiple benchmarks demonstrate that \textsc{UniTE} significantly enhances performance compared to existing methods, offering a more efficient framework for LLM ensembling.

SGW-based Multi-Task Learning in Vision Tasks

Multi-task-learning(MTL) is a multi-target optimization task. Neural networks try to realize each target using a shared interpretative space within MTL. However, as the scale of datasets expands and the complexity of tasks increases, knowledge sharing becomes increasingly challenging. In this paper, we first re-examine previous cross-attention MTL methods from the perspective of noise. We theoretically analyze this issue and identify it as a flaw in the cross-attention mechanism. To address this issue, we propose an information bottleneck knowledge extraction module (KEM). This module aims to reduce inter-task interference by constraining the flow of information, thereby reducing computational complexity. Furthermore, we have employed neural collapse to stabilize the knowledge-selection process. That is, before input to KEM, we projected the features into ETF space. This mapping makes our method more robust. We implemented and conducted comparative experiments with this method on multiple datasets. The results demonstrate that our approach significantly outperforms existing methods in multi-task learning.

Learning Attentional Mixture of LoRAs for Language Model Continual Learning

Fine-tuning large language models (LLMs) with Low-Rank adaption (LoRA) is widely acknowledged as an effective approach for continual learning for new tasks. However, it often suffers from catastrophic forgetting when dealing with multiple tasks sequentially. To this end, we propose Attentional Mixture of LoRAs (AM-LoRA), a continual learning approach tailored for LLMs. Specifically, AM-LoRA learns a sequence of LoRAs for a series of tasks to continually learn knowledge from different tasks. The key of our approach is that we devise an attention mechanism as a knowledge mixture module to adaptively integrate information from each LoRA. With the attention mechanism, AM-LoRA can efficiently leverage the distinctive contributions of each LoRA, while mitigating the risk of mutually negative interactions among them that may lead to catastrophic forgetting. Moreover, we further introduce $L1$ norm in the learning process to make the attention vector more sparse. The sparse constraints can enable the model to lean towards selecting a few highly relevant LoRAs, rather than aggregating and weighting all LoRAs collectively, which can further reduce the impact stemming from mutual interference. Experimental results on continual learning benchmarks indicate the superiority of our proposed method.

DualSep: A Light-weight dual-encoder convolutional recurrent network for real-time in-car speech separation

Advancements in deep learning and voice-activated technologies have driven the development of human-vehicle interaction. Distributed microphone arrays are widely used in in-car scenarios because they can accurately capture the voices of passengers from different speech zones. However, the increase in the number of audio channels, coupled with the limited computational resources and low latency requirements of in-car systems, presents challenges for in-car multi-channel speech separation. To migrate the problems, we propose a lightweight framework that cascades digital signal processing (DSP) and neural networks (NN). We utilize fixed beamforming (BF) to reduce computational costs and independent vector analysis (IVA) to provide spatial prior. We employ dual encoders for dual-branch modeling, with spatial encoder capturing spatial cues and spectral encoder preserving spectral information, facilitating spatial-spectral fusion. Our proposed system supports both streaming and non-streaming modes. Experimental results demonstrate the superiority of the proposed system across various metrics. With only 0.83M parameters and 0.39 real-time factor (RTF) on an Intel Core i7 (2.6GHz) CPU, it effectively separates speech into distinct speech zones. Our demos are available at https://honee-w.github.io/DualSep/.

MarsCode Agent: AI-native Automated Bug Fixing

Recent advances in large language models (LLMs) have shown significant potential to automate various software development tasks, including code completion, test generation, and bug fixing. However, the application of LLMs for automated bug fixing remains challenging due to the complexity and diversity of real-world software systems. In this paper, we introduce MarsCode Agent, a novel framework that leverages LLMs to automatically identify and repair bugs in software code. MarsCode Agent combines the power of LLMs with advanced code analysis techniques to accurately localize faults and generate patches. Our approach follows a systematic process of planning, bug reproduction, fault localization, candidate patch generation, and validation to ensure high-quality bug fixes. We evaluated MarsCode Agent on SWE-bench, a comprehensive benchmark of real-world software projects, and our results show that MarsCode Agent achieves a high success rate in bug fixing compared to most of the existing automated approaches.