Hui Huang

Hui Huang, Xingyuan Bu, Hongli Zhou et al.

Recently, there has been a growing trend of utilizing Large Language Model (LLM) to evaluate the quality of other LLMs. Many studies have fine-tuned judge models based on open-source LLMs for evaluation. While the fine-tuned judge models are claimed to achieve comparable evaluation capability with GPT-4, in this work, we conduct an empirical study of LLM-as-a-Judge. Our findings indicate that although the fine-tuned judge models achieve high performance on in-domain test sets, even surpassing GPT-4, they underperform GPT-4 across several dimensions, including generalizability, fairness and adaptability. We also reveal that the fine-tuned judge model inherently operates as a task-specific classifier, consequently imposing the limitations.

Hui Huang, Jiaheng Liu, Yancheng He et al.

Complex instruction-following with elaborate constraints is imperative for Large Language Models (LLMs). While existing methods have constructed data for complex instruction alignment, they all rely on a more advanced model, especially GPT-4, limiting their application. In this paper, we propose a Multi-granularity Self-Contrastive Training (MuSC) framework, to improve the complex instruction alignment without relying on a stronger model. Our method is conducted on both coarse and fine granularity. On coarse-granularity, we construct constraint-aware preference data based on instruction decomposition and recombination. On fine-granularity, we perform token-aware preference optimization with dynamic token-level supervision. Our method is evaluated on open-sourced models, and experiment results show our method achieves significant improvement on both complex and general instruction-following benchmarks, surpassing previous self-alignment methods.

Wenqi Fang, Huiyun Li, Hui Huang et al.

Gaussian process is one of the most popular non-parametric Bayesian methodologies for modeling the regression problem. It is completely determined by its mean and covariance functions. And its linear property makes it relatively straightforward to solve the prediction problem. Although Gaussian process has been successfully applied in many fields, it is still not enough to deal with physical systems that satisfy inequality constraints. This issue has been addressed by the so-called constrained Gaussian process in recent years. In this paper, we extend the core ideas of constrained Gaussian process. According to the range of training or test data, we redefine the hat basis functions mentioned in the constrained Gaussian process. Based on hat basis functions, we propose a new sparse Gaussian process method to solve the unconstrained regression problem. Similar to the exact Gaussian process and Gaussian process with Fully Independent Training Conditional approximation, our method obtains satisfactory approximate results on open-source datasets or analytical functions. In terms of performance, the proposed method reduces the overall computational complexity from $O(n^{3})$ computation in exact Gaussian process to $O(nm^{2})$ with $m$ hat basis functions and $n$ training data points.

Hang Zhao, Chenyang Zhu, Xin Xu et al.

We tackle the Online 3D Bin Packing Problem, a challenging yet practically useful variant of the classical Bin Packing Problem. In this problem, the items are delivered to the agent without informing the full sequence information. Agent must directly pack these items into the target bin stably without changing their arrival order, and no further adjustment is permitted. Online 3D-BPP can be naturally formulated as Markov Decision Process (MDP). We adopt deep reinforcement learning, in particular, the on-policy actor-critic framework, to solve this MDP with constrained action space. To learn a practically feasible packing policy, we propose three critical designs. First, we propose an online analysis of packing stability based on a novel stacking tree. It attains a high analysis accuracy while reducing the computational complexity from $O(N^2)$ to $O(N \log N)$, making it especially suited for RL training. Second, we propose a decoupled packing policy learning for different dimensions of placement which enables high-resolution spatial discretization and hence high packing precision. Third, we introduce a reward function that dictates the robot to place items in a far-to-near order and therefore simplifies the collision avoidance in movement planning of the robotic arm. Furthermore, we provide a comprehensive discussion on several key implemental issues. The extensive evaluation demonstrates that our learned policy outperforms the state-of-the-art methods significantly and is practically usable for real-world applications.