Reward Modeling with Ordinal Feedback: Wisdom of the Crowd

Learning a reward model (RM) from human preferences has been an important component in aligning large language models (LLMs). The canonical setup of learning RMs from pairwise preference data is rooted in the classic Bradley-Terry (BT) model that accepts binary feedback, i.e., the label being either Response 1 is better than Response 2, or the opposite. Such a setup inevitably discards potentially useful samples (such as "tied" between the two responses) and loses more fine-grained information (such as "slightly better"). In this paper, we propose a framework for learning RMs under ordinal feedback which generalizes the case of binary preference feedback to any arbitrary granularity. Specifically, we first identify a marginal unbiasedness condition, which generalizes the assumption of the BT model in the existing binary feedback setting. The condition validates itself via the sociological concept of the wisdom of the crowd. Under the condition, we develop a natural probability model for pairwise preference data under ordinal feedback and analyze its properties. We prove the statistical benefits of ordinal feedback in terms of reducing the Rademacher complexity compared to the case of binary feedback. The proposed learning objective and the theory also extend to hinge loss and direct policy optimization (DPO). In particular, the theoretical analysis may be of independent interest when applying to a seemingly unrelated problem of knowledge distillation to interpret the bias-variance trade-off therein. The framework also sheds light on writing guidance for human annotators. Our numerical experiments validate that fine-grained feedback leads to better reward learning for both in-distribution and out-of-distribution settings. Further experiments show that incorporating a certain proportion of samples with tied preference boosts RM learning.

Fire-Flyer AI-HPC: A Cost-Effective Software-Hardware Co-Design for Deep Learning

The rapid progress in Deep Learning (DL) and Large Language Models (LLMs) has exponentially increased demands of computational power and bandwidth. This, combined with the high costs of faster computing chips and interconnects, has significantly inflated High Performance Computing (HPC) construction costs. To address these challenges, we introduce the Fire-Flyer AI-HPC architecture, a synergistic hardware-software co-design framework and its best practices. For DL training, we deployed the Fire-Flyer 2 with 10,000 PCIe A100 GPUs, achieved performance approximating the DGX-A100 while reducing costs by half and energy consumption by 40%. We specifically engineered HFReduce to accelerate allreduce communication and implemented numerous measures to keep our Computation-Storage Integrated Network congestion-free. Through our software stack, including HaiScale, 3FS, and HAI-Platform, we achieved substantial scalability by overlapping computation and communication. Our system-oriented experience from DL training provides valuable insights to drive future advancements in AI-HPC.

Towards Better Understanding of In-Context Learning Ability from In-Context Uncertainty Quantification

Predicting simple function classes has been widely used as a testbed for developing theory and understanding of the trained Transformer's in-context learning (ICL) ability. In this paper, we revisit the training of Transformers on linear regression tasks, and different from all the existing literature, we consider a bi-objective prediction task of predicting both the conditional expectation $\mathbb{E}[Y|X]$ and the conditional variance Var$(Y|X)$. This additional uncertainty quantification objective provides a handle to (i) better design out-of-distribution experiments to distinguish ICL from in-weight learning (IWL) and (ii) make a better separation between the algorithms with and without using the prior information of the training distribution. Theoretically, we show that the trained Transformer reaches near Bayes-optimum, suggesting the usage of the information of the training distribution. Our method can be extended to other cases. Specifically, with the Transformer's context window $S$, we prove a generalization bound of $\tilde{\mathcal{O}}(\sqrt{\min\{S, T\}/(n T)})$ on $n$ tasks with sequences of length $T$, providing sharper analysis compared to previous results of $\tilde{\mathcal{O}}(\sqrt{1/n})$. Empirically, we illustrate that while the trained Transformer behaves as the Bayes-optimal solution as a natural consequence of supervised training in distribution, it does not necessarily perform a Bayesian inference when facing task shifts, in contrast to the \textit{equivalence} between these two proposed in many existing literature. We also demonstrate the trained Transformer's ICL ability over covariates shift and prompt-length shift and interpret them as a generalization over a meta distribution.

Understanding the Training and Generalization of Pretrained Transformer for Sequential Decision Making

In this paper, we consider the supervised pretrained transformer for a class of sequential decision-making problems. The class of considered problems is a subset of the general formulation of reinforcement learning in that there is no transition probability matrix, and the class of problems covers bandits, dynamic pricing, and newsvendor problems as special cases. Such a structure enables the use of optimal actions/decisions in the pretraining phase, and the usage also provides new insights for the training and generalization of the pretrained transformer. We first note that the training of the transformer model can be viewed as a performative prediction problem, and the existing methods and theories largely ignore or cannot resolve the arisen out-of-distribution issue. We propose a natural solution that includes the transformer-generated action sequences in the training procedure, and it enjoys better properties both numerically and theoretically. The availability of the optimal actions in the considered tasks also allows us to analyze the properties of the pretrained transformer as an algorithm and explains why it may lack exploration and how this can be automatically resolved. Numerically, we categorize the advantages of the pretrained transformer over the structured algorithms such as UCB and Thompson sampling into three cases: (i) it better utilizes the prior knowledge in the pretraining data; (ii) it can elegantly handle the misspecification issue suffered by the structured algorithms; (iii) for short time horizon such as $T\le50$, it behaves more greedy and enjoys much better regret than the structured algorithms which are designed for asymptotic optimality.

Uncertainty Estimation and Quantification for LLMs: A Simple Supervised Approach

Large language models (LLMs) are highly capable of many tasks but they can sometimes generate unreliable or inaccurate outputs. To tackle this issue, this paper studies the problem of uncertainty estimation and calibration for LLMs. We begin by formulating the uncertainty estimation problem for LLMs and then propose a supervised approach that takes advantage of the labeled datasets and estimates the uncertainty of the LLMs' responses. Based on the formulation, we illustrate the difference between the uncertainty estimation for LLMs and that for standard ML models and explain why the hidden activations of the LLMs contain uncertainty information. Our designed approach effectively demonstrates the benefits of utilizing hidden activations for enhanced uncertainty estimation across various tasks and shows robust transferability in out-of-distribution settings. Moreover, we distinguish the uncertainty estimation task from the uncertainty calibration task and show that a better uncertainty estimation mode leads to a better calibration performance. In practice, our method is easy to implement and is adaptable to different levels of model transparency including black box, grey box, and white box, each demonstrating strong performance based on the accessibility of the LLM's internal mechanisms.

DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence

The rapid development of large language models has revolutionized code intelligence in software development. However, the predominance of closed-source models has restricted extensive research and development. To address this, we introduce the DeepSeek-Coder series, a range of open-source code models with sizes from 1.3B to 33B, trained from scratch on 2 trillion tokens. These models are pre-trained on a high-quality project-level code corpus and employ a fill-in-the-blank task with a 16K window to enhance code generation and infilling. Our extensive evaluations demonstrate that DeepSeek-Coder not only achieves state-of-the-art performance among open-source code models across multiple benchmarks but also surpasses existing closed-source models like Codex and GPT-3.5. Furthermore, DeepSeek-Coder models are under a permissive license that allows for both research and unrestricted commercial use.

DeepSeek LLM: Scaling Open-Source Language Models with Longtermism

The rapid development of open-source large language models (LLMs) has been truly remarkable. However, the scaling law described in previous literature presents varying conclusions, which casts a dark cloud over scaling LLMs. We delve into the study of scaling laws and present our distinctive findings that facilitate scaling of large scale models in two commonly used open-source configurations, 7B and 67B. Guided by the scaling laws, we introduce DeepSeek LLM, a project dedicated to advancing open-source language models with a long-term perspective. To support the pre-training phase, we have developed a dataset that currently consists of 2 trillion tokens and is continuously expanding. We further conduct supervised fine-tuning (SFT) and Direct Preference Optimization (DPO) on DeepSeek LLM Base models, resulting in the creation of DeepSeek Chat models. Our evaluation results demonstrate that DeepSeek LLM 67B surpasses LLaMA-2 70B on various benchmarks, particularly in the domains of code, mathematics, and reasoning. Furthermore, open-ended evaluations reveal that DeepSeek LLM 67B Chat exhibits superior performance compared to GPT-3.5.

Learning to Make Adherence-Aware Advice

As artificial intelligence (AI) systems play an increasingly prominent role in human decision-making, challenges surface in the realm of human-AI interactions. One challenge arises from the suboptimal AI policies due to the inadequate consideration of humans disregarding AI recommendations, as well as the need for AI to provide advice selectively when it is most pertinent. This paper presents a sequential decision-making model that (i) takes into account the human's adherence level (the probability that the human follows/rejects machine advice) and (ii) incorporates a defer option so that the machine can temporarily refrain from making advice. We provide learning algorithms that learn the optimal advice policy and make advice only at critical time stamps. Compared to problem-agnostic reinforcement learning algorithms, our specialized learning algorithms not only enjoy better theoretical convergence properties but also show strong empirical performance.

Facilitating Battery Swapping Services for Freight Trucks with Spatial-Temporal Demand Prediction

Electrifying heavy-duty trucks offers a substantial opportunity to curtail carbon emissions, advancing toward a carbon-neutral future. However, the inherent challenges of limited battery energy and the sheer weight of heavy-duty trucks lead to reduced mileage and prolonged charging durations. Consequently, battery-swapping services emerge as an attractive solution for these trucks. This paper employs a two-fold approach to investigate the potential and enhance the efficacy of such services. Firstly, spatial-temporal demand prediction models are adopted to predict the traffic patterns for the upcoming hours. Subsequently, the prediction guides an optimization module for efficient battery allocation and deployment. Analyzing the heavy-duty truck data on a highway network spanning over 2,500 miles, our model and analysis underscore the value of prediction/machine learning in facilitating future decision-makings. In particular, we find that the initial phase of implementing battery-swapping services favors mobile battery-swapping stations, but as the system matures, fixed-location stations are preferred.

Fairer LP-based Online Allocation

In this paper, we consider a Linear Program (LP)-based online resource allocation problem where a decision maker accepts or rejects incoming customer requests irrevocably in order to maximize expected revenue given limited resources. At each time, a new order/customer/bid is revealed with a request of some resource(s) and a reward. We consider a stochastic setting where all the orders are i.i.d. sampled from an unknown distribution. Such formulation gives rise to many classic applications such as the canonical (quantity-based) network revenue management problem and the Adwords problem. Instead of focusing only on regret minimization, this paper aims to provide fairness guarantees while maintaining low regret. Our definition of fairness is that a fair online algorithm should treat similar agents/customers similarly and the decision made for similar individuals should be consistent over time. We define the fair offline solution as the analytic center of the offline optimal solution set, and define \textit{cumulative unfairness} as the cumulative deviation from the online solutions to the fair offline solution. We propose a fair algorithm that uses an interior-point LP solver and dynamically detects unfair resource spending. Our algorithm can control cumulative unfairness on the scale of order $O(\log(T))$, while maintaining the regret to be bounded without dependency on $T$. Moreover, we partially remove the nondegeneracy assumptions used in early results in the literature. This paper only requires the nondegeneracy condition for the binding constraints, and allows the existence of multiple optimal solutions.