Xingyuan Pan

Xian Qian, Kai Hu, Jiaqiang Wang et al. · bytedance

This report describes our VolcTrans system for the WMT22 shared task on large-scale multilingual machine translation. We participated in the unconstrained track which allows the use of external resources. Our system is a transformerbased multilingual model trained on data from multiple sources including the public training set from the data track, NLLB data provided by Meta AI, self-collected parallel corpora, and pseudo bitext from back-translation. A series of heuristic rules clean both bilingual and monolingual texts. On the official test set, our system achieves 17.3 BLEU, 21.9 spBLEU, and 41.9 chrF2++ on average over all language pairs. The average inference speed is 11.5 sentences per second using a single Nvidia Tesla V100 GPU. Our code and trained models are available at https://github.com/xian8/wmt22

Rui Pan, Shivanshu Shekhar, Boyao Wang et al.

Small language models (SLMs) have attracted considerable attention from both academia and industry due to their broad range of applications in edge devices. To obtain SLMs with strong performance, conventional approaches either pre-train the models from scratch, which incurs substantial computational costs, or compress/prune existing large language models (LLMs), which results in performance drops and falls short in comparison to pre-training. In this paper, we investigate the family of acceleration methods that involve both structured pruning and model training. We found 1) layer-wise adaptive pruning (Adapt-Pruner) is extremely effective in LLMs and yields significant improvements over existing pruning techniques, 2) adaptive pruning equipped with further training leads to models comparable to those pre-training from scratch, 3) incremental pruning brings non-trivial performance gain by interleaving pruning with training and only removing a small portion of neurons ($\sim$5%) at a time. Experimental results on LLaMA-3.1-8B demonstrate that Adapt-Pruner outperforms conventional pruning methods, such as LLM-Pruner, FLAP, and SliceGPT, by an average of 1%-7% in accuracy on commonsense benchmarks. Additionally, Adapt-Pruner restores the performance of MobileLLM-125M to 600M on the MMLU benchmark with 200$\times$ fewer tokens via pruning from its larger counterparts, and discovers a new 1B model that surpasses LLaMA-3.2-1B in multiple benchmarks. The official code is released at https://github.com/research4pan/AdaptPruner.

Guangrui Li, Yaochen Xie, Yi Liu et al.

LLM-powered agents fulfill user requests by interacting with environments, querying data, and invoking tools in a multi-turn process. Yet, most existing benchmarks assume static environments with fixed schemas and toolsets, neglecting the evolutionary nature of the real world and agents' robustness to environmental changes. In this paper, we study a crucial problem: how to evolve the agent environment in a scalable and controllable way, thereby better evaluating agents' adaptability to real-world dynamics. We propose ProEvolve, a graph-based framework that makes environment evolution programmable. At its core, a typed relational graph provides a unified, explicit representation of the environment: data, tools, and schema. Under this formalism, adding, removing, or modifying capabilities are expressed as graph transformations that coherently propagate updates across tools, schemas, and data access. Building on this, ProEvolve can (1) program the evolutionary dynamics as graph transformations to generate environments automatically, and (2) instantiate task sandboxes via subgraph sampling and programming. We validate ProEvolve by evolving a single environment into 200 environments and 3,000 task sandboxes, and benchmark representative agents accordingly.

Xingyuan Pan, Luyang Huang, Liyan Kang et al.

Large Language Models (LLMs) have demonstrated remarkable abilities in general scenarios. Instruction finetuning empowers them to align with humans in various tasks. Nevertheless, the Diversity and Quality of the instruction data remain two main challenges for instruction finetuning. With regard to this, in this paper, we propose a novel gradient-based method to automatically select high-quality and diverse instruction finetuning data for machine translation. Our key innovation centers around analyzing how individual training examples influence the model during training. Specifically, we select training examples that exert beneficial influences on the model as high-quality ones by means of Influence Function plus a small high-quality seed dataset. Moreover, to enhance the diversity of the training data we maximize the variety of influences they have on the model by clustering on their gradients and resampling. Extensive experiments on WMT22 and FLORES translation tasks demonstrate the superiority of our methods, and in-depth analysis further validates their effectiveness and generalization.

Yifan Hao, Xingyuan Pan, Hanning Zhang et al.

Supervised fine-tuning (SFT) on domain-specific data is the dominant approach for adapting foundation models to specialized tasks. However, it has been observed that SFT models tend to forget knowledge acquired during pretraining. In vision models, ensembling a pretrained model with its fine-tuned counterpart has been shown to mitigate this issue. In this work, we demonstrate that the same holds for language models, and, more strikingly, we observe an overadaptation phenomenon: the ensemble model not only retains general knowledge from the foundation model but also outperforms the fine-tuned model even on the fine-tuning domain itself. Despite the empirical success of ensembling, a theoretical understanding of its benefits remains underexplored. We develop a formal theoretical analysis of the overadaptation phenomenon. Ensembling mitigates this by balancing two primary sources of error: bias, caused by insufficient fine-tuning, and variance, introduced by overfitting to fine-tuning data. While regularization techniques aim to address this trade-off, we show that ensembling provides a more effective solution. We analyze this phenomenon in over-parameterized linear settings and demonstrate that interpolating between pretrained and fine-tuned weights significantly improves performance. These findings offer theoretical justification for the observed advantages of model ensembling, supported by empirical experiments consistent with our analysis.

Xingyuan Pan, Chenlu Ye, Joseph Melkonian et al.

Training data attribution (TDA) methods aim to identify which training examples influence a model's predictions on specific test data most. By quantifying these influences, TDA supports critical applications such as data debugging, curation, and valuation. Gradient-based TDA methods rely on gradients and second-order information, limiting their applicability at scale. While recent random projection-based methods improve scalability, they often suffer from degraded attribution accuracy. Motivated by connections between uncertainty and influence functions, we introduce Daunce - a simple yet effective data attribution approach through uncertainty estimation. Our method operates by fine-tuning a collection of perturbed models and computing the covariance of per-example losses across these models as the attribution score. Daunce is scalable to large language models (LLMs) and achieves more accurate attribution compared to existing TDA methods. We validate Daunce on tasks ranging from vision tasks to LLM fine-tuning, and further demonstrate its compatibility with black-box model access. Applied to OpenAI's GPT models, our method achieves, to our knowledge, the first instance of data attribution on proprietary LLMs.

Rui Pan, Dylan Zhang, Hanning Zhang et al.

Bilevel optimization has shown its utility across various machine learning settings, yet most algorithms in practice require second-order information, making it challenging to scale them up. Only recently, a paradigm of first-order algorithms has emerged in the theoretical literature, capable of effectively addressing bilevel optimization problems. Nevertheless, the practical efficiency of this paradigm remains unverified, particularly in the context of large language models (LLMs). This paper introduces the first scalable instantiation of this paradigm called ScaleBiO, focusing on bilevel optimization for large-scale LLM data reweighting. By combining with a recently proposed memory-efficient training technique called LISA, our novel algorithm allows the paradigm to scale to $\sim$30B-sized LLMs on $8\times$H100 GPUs, marking the first successful application of bilevel optimization under practical scenarios for large-sized LLMs. Empirically, extensive experiments on data reweighting verify the effectiveness of ScaleBiO for different-scaled models, including Llama-3-8B, Gemma-2-9B, Qwen-2-7B, and Qwen-2.5-32B, where bilevel optimization succeeds in instruction-following and math reasoning tasks, outperforming several popular baselines, including uniform sampling, influence-aware data filtering, and reference-model-based sampling methods. Theoretically, ScaleBiO ensures the optimality of the learned data weights, along with a convergence guarantee matching the conventional first-order bilevel optimization paradigm on smooth and strongly convex objectives.

Xingyuan Pan, Maitrey Mehta, Vivek Srikumar

Various natural language processing tasks are structured prediction problems where outputs are constructed with multiple interdependent decisions. Past work has shown that domain knowledge, framed as constraints over the output space, can help improve predictive accuracy. However, designing good constraints often relies on domain expertise. In this paper, we study the problem of learning such constraints. We frame the problem as that of training a two-layer rectifier network to identify valid structures or substructures, and show a construction for converting a trained network into a system of linear constraints over the inference variables. Our experiments on several NLP tasks show that the learned constraints can improve the prediction accuracy, especially when the number of training examples is small.

Xingyuan Pan, Vivek Srikumar

Predicting structured outputs can be computationally onerous due to the combinatorially large output spaces. In this paper, we focus on reducing the prediction time of a trained black-box structured classifier without losing accuracy. To do so, we train a speedup classifier that learns to mimic a black-box classifier under the learning-to-search approach. As the structured classifier predicts more examples, the speedup classifier will operate as a learned heuristic to guide search to favorable regions of the output space. We present a mistake bound for the speedup classifier and identify inference situations where it can independently make correct judgments without input features. We evaluate our method on the task of entity and relation extraction and show that the speedup classifier outperforms even greedy search in terms of speed without loss of accuracy.

Xingyuan Pan, Vivek Srikumar

Rectified Linear Units (ReLUs) have been shown to ameliorate the vanishing gradient problem, allow for efficient backpropagation, and empirically promote sparsity in the learned parameters. They have led to state-of-the-art results in a variety of applications. However, unlike threshold and sigmoid networks, ReLU networks are less explored from the perspective of their expressiveness. This paper studies the expressiveness of ReLU networks. We characterize the decision boundary of two-layer ReLU networks by constructing functionally equivalent threshold networks. We show that while the decision boundary of a two-layer ReLU network can be captured by a threshold network, the latter may require an exponentially larger number of hidden units. We also formulate sufficient conditions for a corresponding logarithmic reduction in the number of hidden units to represent a sign network as a ReLU network. Finally, we experimentally compare threshold networks and their much smaller ReLU counterparts with respect to their ability to learn from synthetically generated data.