A Practice of Post-Training on Llama-3 70B with Optimal Selection of Additional Language Mixture Ratio

Large Language Models (LLM) often needs to be Continual Pre-Trained (CPT) to obtain the unfamiliar language skill or adapt into new domains. The huge training cost of CPT often asks for cautious choice of key hyper-parameters such as the mixture ratio of extra language or domain corpus. However, there is no systematic study which bridge the gap between the optimal mixture ratio and the actual model performance, and the gap between experimental scaling law and the actual deployment in the full model size. In this paper, we perform CPT on Llama-3 8B and 70B to enhance its Chinese ability. We study the optimal correlation between the Additional Language Mixture Ratio (ALMR) and the Learning Rate (LR) on the 8B size which directly indicate the optimal experimental set up. By thorough choice of hyper-parameter, and subsequent fine-tuning, the model capability is improved not only on the Chinese-related benchmark, but also some specific domains including math, coding and emotional intelligence. We deploy the final 70B version of LLM on an real-life chat system which obtain satisfying performance.

LTNER: Large Language Model Tagging for Named Entity Recognition with Contextualized Entity Marking

The use of LLMs for natural language processing has become a popular trend in the past two years, driven by their formidable capacity for context comprehension and learning, which has inspired a wave of research from academics and industry professionals. However, for certain NLP tasks, such as NER, the performance of LLMs still falls short when compared to supervised learning methods. In our research, we developed a NER processing framework called LTNER that incorporates a revolutionary Contextualized Entity Marking Gen Method. By leveraging the cost-effective GPT-3.5 coupled with context learning that does not require additional training, we significantly improved the accuracy of LLMs in handling NER tasks. The F1 score on the CoNLL03 dataset increased from the initial 85.9% to 91.9%, approaching the performance of supervised fine-tuning. This outcome has led to a deeper understanding of the potential of LLMs.

Leveraging Large Language Model for Automatic Evolving of Industrial Data-Centric R&D Cycle

In the wake of relentless digital transformation, data-driven solutions are emerging as powerful tools to address multifarious industrial tasks such as forecasting, anomaly detection, planning, and even complex decision-making. Although data-centric R&D has been pivotal in harnessing these solutions, it often comes with significant costs in terms of human, computational, and time resources. This paper delves into the potential of large language models (LLMs) to expedite the evolution cycle of data-centric R&D. Assessing the foundational elements of data-centric R&D, including heterogeneous task-related data, multi-facet domain knowledge, and diverse computing-functional tools, we explore how well LLMs can understand domain-specific requirements, generate professional ideas, utilize domain-specific tools to conduct experiments, interpret results, and incorporate knowledge from past endeavors to tackle new challenges. We take quantitative investment research as a typical example of industrial data-centric R&D scenario and verified our proposed framework upon our full-stack open-sourced quantitative research platform Qlib and obtained promising results which shed light on our vision of automatic evolving of industrial data-centric R&D cycle.

Linear TreeShap

Decision trees are well-known due to their ease of interpretability. To improve accuracy, we need to grow deep trees or ensembles of trees. These are hard to interpret, offsetting their original benefits. Shapley values have recently become a popular way to explain the predictions of tree-based machine learning models. It provides a linear weighting to features independent of the tree structure. The rise in popularity is mainly due to TreeShap, which solves a general exponential complexity problem in polynomial time. Following extensive adoption in the industry, more efficient algorithms are required. This paper presents a more efficient and straightforward algorithm: Linear TreeShap. Like TreeShap, Linear TreeShap is exact and requires the same amount of memory.

On a scalable problem transformation method for multi-label learning

Binary relevance is a simple approach to solve multi-label learning problems where an independent binary classifier is built per each label. A common challenge with this in real-world applications is that the label space can be very large, making it difficult to use binary relevance to larger scale problems. In this paper, we propose a scalable alternative to this, via transforming the multi-label problem into a single binary classification. We experiment with a few variations of our method and show that our method achieves higher precision than binary relevance and faster execution times on a top-K recommender system task.

Negative Log Likelihood Ratio Loss for Deep Neural Network Classification

In deep neural network, the cross-entropy loss function is commonly used for classification. Minimizing cross-entropy is equivalent to maximizing likelihood under assumptions of uniform feature and class distributions. It belongs to generative training criteria which does not directly discriminate correct class from competing classes. We propose a discriminative loss function with negative log likelihood ratio between correct and competing classes. It significantly outperforms the cross-entropy loss on the CIFAR-10 image classification task.

Time Resource Networks

The problem of scheduling under resource constraints is widely applicable. One prominent example is power management, in which we have a limited continuous supply of power but must schedule a number of power-consuming tasks. Such problems feature tightly coupled continuous resource constraints and continuous temporal constraints. We address such problems by introducing the Time Resource Network (TRN), an encoding for resource-constrained scheduling problems. The definition allows temporal specifications using a general family of representations derived from the Simple Temporal network, including the Simple Temporal Network with Uncertainty, and the probabilistic Simple Temporal Network (Fang et al. (2014)). We propose two algorithms for determining the consistency of a TRN: one based on Mixed Integer Programing and the other one based on Constraint Programming, which we evaluate on scheduling problems with Simple Temporal Constraints and Probabilistic Temporal Constraints.