OCEAN: Offline Chain-of-thought Evaluation and Alignment in Large Language Models

Offline evaluation of LLMs is crucial in understanding their capacities, though current methods remain underexplored in existing research. In this work, we focus on the offline evaluation of the chain-of-thought capabilities and show how to optimize LLMs based on the proposed evaluation method. To enable offline feedback with rich knowledge and reasoning paths, we use knowledge graphs (e.g., Wikidata5m) to provide feedback on the generated chain of thoughts. Due to the heterogeneity between LLM reasoning and KG structures, direct interaction and feedback from KGs on LLM behavior are challenging, as they require accurate entity linking and grounding of LLM-generated chains of thought in the KG. To address the above challenge, we propose an offline chain-of-thought evaluation framework, OCEAN, which models chain-of-thought reasoning in LLMs as an MDP and evaluate the policy's alignment with KG preference modeling. To overcome the reasoning heterogeneity and grounding problems, we leverage on-policy KG exploration and RL to model a KG policy that generates token-level likelihood distributions for LLM-generated chain-of-thought reasoning paths, simulating KG reasoning preference. Then we incorporate the knowledge-graph feedback on the validity and alignment of the generated reasoning paths into inverse propensity scores and propose KG-IPS estimator. Theoretically, we prove the unbiasedness of the proposed KG-IPS estimator and provide a lower bound on its variance. With the off-policy evaluated value function, we can directly enable off-policy optimization to further enhance chain-of-thought alignment. Our empirical study shows that OCEAN can be efficiently optimized for generating chain-of-thought reasoning paths with higher estimated values without affecting LLMs' general abilities in downstream tasks or their internal knowledge.

Augmenting the Veracity and Explanations of Complex Fact Checking via Iterative Self-Revision with LLMs

Explanation generation plays a more pivotal role than fact verification in producing interpretable results and facilitating comprehensive fact-checking, which has recently garnered considerable attention. However, previous studies on explanation generation has shown several limitations, such as being confined to English scenarios, involving overly complex inference processes, and not fully unleashing the potential of the mutual feedback between veracity labels and explanation texts. To address these issues, we construct two complex fact-checking datasets in the Chinese scenarios: CHEF-EG and TrendFact. These datasets involve complex facts in areas such as health, politics, and society, presenting significant challenges for fact verification methods. In response to these challenges, we propose a unified framework called FactISR (Augmenting Fact-Checking via Iterative Self-Revision) to perform mutual feedback between veracity and explanations by leveraging the capabilities of large language models(LLMs). FactISR uses a single model to address tasks such as fact verification and explanation generation. Its self-revision mechanism can further revision the consistency between veracity labels, explanation texts, and evidence, as well as eliminate irrelevant noise. We conducted extensive experiments with baselines and FactISR on the proposed datasets. The experimental results demonstrate the effectiveness of our method.

Self-Evolutionary Large Language Models through Uncertainty-Enhanced Preference Optimization

Iterative preference optimization has recently become one of the de-facto training paradigms for large language models (LLMs), but the performance is still underwhelming due to too much noisy preference data yielded in the loop. To combat this issue, we present an \textbf{U}ncertainty-enhanced \textbf{P}reference \textbf{O}ptimization (UPO) framework to make the LLM self-evolve with reliable feedback. The key idea is mitigating the noisy preference data derived from the current policy and reward models by performing pair-wise uncertainty estimation and judiciously reliable feedback sampling. To reach this goal, we thus introduce an estimator model, which incorporates Monte Carlo (MC) dropout in Bayesian neural network (BNN) to perform uncertainty estimation for the preference data derived from the LLM policy. Compared to the existing methods that directly filter generated responses based on the reward score, the estimator focuses on the model uncertainty in a pair-wise manner and effectively bypasses the confirmation bias problem of the reward model. Additionally, we also propose an uncertainty-enhanced self-evolution algorithm to improve the robustness of preference optimization and encourage the LLM to generate responses with both high reward and certainty. Extensive experiments over multiple benchmarks demonstrate that our framework substantially alleviates the noisy problem and improves the performance of iterative preference optimization.

Causal Deepsets for Off-policy Evaluation under Spatial or Spatio-temporal Interferences

Off-policy evaluation (OPE) is widely applied in sectors such as pharmaceuticals and e-commerce to evaluate the efficacy of novel products or policies from offline datasets. This paper introduces a causal deepset framework that relaxes several key structural assumptions, primarily the mean-field assumption, prevalent in existing OPE methodologies that handle spatio-temporal interference. These traditional assumptions frequently prove inadequate in real-world settings, thereby restricting the capability of current OPE methods to effectively address complex interference effects. In response, we advocate for the implementation of the permutation invariance (PI) assumption. This innovative approach enables the data-driven, adaptive learning of the mean-field function, offering a more flexible estimation method beyond conventional averaging. Furthermore, we present novel algorithms that incorporate the PI assumption into OPE and thoroughly examine their theoretical foundations. Our numerical analyses demonstrate that this novel approach yields significantly more precise estimations than existing baseline algorithms, thereby substantially improving the practical applicability and effectiveness of OPE methodologies. A Python implementation of our proposed method is available at https://github.com/BIG-S2/Causal-Deepsets.

A Survey of Neural Code Intelligence: Paradigms, Advances and Beyond

Neural Code Intelligence -- leveraging deep learning to understand, generate, and optimize code -- holds immense potential for transformative impacts on the whole society. Bridging the gap between Natural Language and Programming Language, this domain has drawn significant attention from researchers in both research communities over the past few years. This survey presents a systematic and chronological review of the advancements in code intelligence, encompassing over 50 representative models and their variants, more than 20 categories of tasks, and an extensive coverage of over 680 related works. We follow the historical progression to trace the paradigm shifts across different research phases (e.g., from modeling code with recurrent neural networks to the era of Large Language Models). Concurrently, we highlight the major technical transitions in models, tasks, and evaluations spanning through different stages. For applications, we also observe a co-evolving shift. It spans from initial endeavors to tackling specific scenarios, through exploring a diverse array of tasks during its rapid expansion, to currently focusing on tackling increasingly complex and varied real-world challenges. Building on our examination of the developmental trajectories, we further investigate the emerging synergies between code intelligence and broader machine intelligence, uncovering new cross-domain opportunities and illustrating the substantial influence of code intelligence across various domains. Finally, we delve into both the opportunities and challenges associated with this field, alongside elucidating our insights on the most promising research directions. An ongoing, dynamically updated project and resources associated with this survey have been released at https://github.com/QiushiSun/NCISurvey.

CoRAL: Collaborative Retrieval-Augmented Large Language Models Improve Long-tail Recommendation

The long-tail recommendation is a challenging task for traditional recommender systems, due to data sparsity and data imbalance issues. The recent development of large language models (LLMs) has shown their abilities in complex reasoning, which can help to deduce users' preferences based on very few previous interactions. However, since most LLM-based systems rely on items' semantic meaning as the sole evidence for reasoning, the collaborative information of user-item interactions is neglected, which can cause the LLM's reasoning to be misaligned with task-specific collaborative information of the dataset. To further align LLMs' reasoning to task-specific user-item interaction knowledge, we introduce collaborative retrieval-augmented LLMs, CoRAL, which directly incorporate collaborative evidence into the prompts. Based on the retrieved user-item interactions, the LLM can analyze shared and distinct preferences among users, and summarize the patterns indicating which types of users would be attracted by certain items. The retrieved collaborative evidence prompts the LLM to align its reasoning with the user-item interaction patterns in the dataset. However, since the capacity of the input prompt is limited, finding the minimally-sufficient collaborative information for recommendation tasks can be challenging. We propose to find the optimal interaction set through a sequential decision-making process and develop a retrieval policy learned through a reinforcement learning (RL) framework, CoRAL. Our experimental results show that CoRAL can significantly improve LLMs' reasoning abilities on specific recommendation tasks. Our analysis also reveals that CoRAL can more efficiently explore collaborative information through reinforcement learning.

InstructGraph: Boosting Large Language Models via Graph-centric Instruction Tuning and Preference Alignment

Do current large language models (LLMs) better solve graph reasoning and generation tasks with parameter updates? In this paper, we propose InstructGraph, a framework that empowers LLMs with the abilities of graph reasoning and generation by instruction tuning and preference alignment. Specifically, we first propose a structured format verbalizer to unify all graph data into a universal code-like format, which can simply represent the graph without any external graph-specific encoders. Furthermore, a graph instruction tuning stage is introduced to guide LLMs in solving graph reasoning and generation tasks. Finally, we identify potential hallucination problems in graph tasks and sample negative instances for preference alignment, the target of which is to enhance the output's reliability of the model. Extensive experiments across multiple graph-centric tasks exhibit that InstructGraph can achieve the best performance and outperform GPT-4 and LLaMA2 by more than 13\% and 38\%, respectively.

Prompting Large Language Models with Chain-of-Thought for Few-Shot Knowledge Base Question Generation

The task of Question Generation over Knowledge Bases (KBQG) aims to convert a logical form into a natural language question. For the sake of expensive cost of large-scale question annotation, the methods of KBQG under low-resource scenarios urgently need to be developed. However, current methods heavily rely on annotated data for fine-tuning, which is not well-suited for few-shot question generation. The emergence of Large Language Models (LLMs) has shown their impressive generalization ability in few-shot tasks. Inspired by Chain-of-Thought (CoT) prompting, which is an in-context learning strategy for reasoning, we formulate KBQG task as a reasoning problem, where the generation of a complete question is splitted into a series of sub-question generation. Our proposed prompting method KQG-CoT first retrieves supportive logical forms from the unlabeled data pool taking account of the characteristics of the logical form. Then, we write a prompt to explicit the reasoning chain of generating complicated questions based on the selected demonstrations. To further ensure prompt quality, we extend KQG-CoT into KQG-CoT+ via sorting the logical forms by their complexity. We conduct extensive experiments over three public KBQG datasets. The results demonstrate that our prompting method consistently outperforms other prompting baselines on the evaluated datasets. Remarkably, our KQG-CoT+ method could surpass existing few-shot SoTA results of the PathQuestions dataset by 18.25, 10.72, and 10.18 absolute points on BLEU-4, METEOR, and ROUGE-L, respectively.

Uncertainty-aware Parameter-Efficient Self-training for Semi-supervised Language Understanding

The recent success of large pre-trained language models (PLMs) heavily hinges on massive labeled data, which typically produces inferior performance in low-resource scenarios. To remedy this dilemma, we study self-training as one of the predominant semi-supervised learning (SSL) approaches, which utilizes large-scale unlabeled data to generate synthetic examples. However, too many noisy labels will hurt the model performance, and the self-training procedure requires multiple training iterations making it more expensive if all the model parameters of the PLM are updated. This paper presents UPET, a novel Uncertainty-aware Parameter-Efficient self-Training framework to effectively and efficiently address the labeled data scarcity issue. Specifically, we incorporate Monte Carlo (MC) dropout in Bayesian neural network (BNN) to perform uncertainty estimation for the teacher model and then judiciously select reliable pseudo-labeled examples based on confidence and certainty. During the student training, we introduce multiple parameter-efficient learning (PEL) paradigms that allow the optimization of only a small percentage of parameters. We also propose a novel Easy-Hard Contrastive Tuning to enhance the robustness and generalization. Extensive experiments over multiple downstream tasks demonstrate that UPET achieves a substantial improvement in terms of performance and efficiency. Our codes and data are released at https: //github.com/wjn1996/UPET.

Boosting In-Context Learning with Factual Knowledge

In-Context Learning (ICL) over Large language models (LLMs) aims at solving previously unseen tasks by conditioning on a few training examples, eliminating the need for parameter updates and achieving competitive performance. In this paper, we demonstrate that factual knowledge is imperative for the performance of ICL in three core facets, i.e., the inherent knowledge learned in LLMs, the factual knowledge derived from the selected in-context examples, and the knowledge biases in LLMs for output generation. To unleash the power of LLMs in few-shot learning scenarios, we introduce a novel Knowledgeable In-Context Tuning (KICT) framework to further improve the performance of ICL: 1) injecting factual knowledge to LLMs during continual self-supervised pre-training, 2) judiciously selecting the examples with high knowledge relevance, and 3) calibrating the prediction results based on prior knowledge. We evaluate the proposed approaches on auto-regressive LLMs (e.g., GPT-style models) over multiple text classification and question answering tasks. Experimental results demonstrate that KICT substantially outperforms strong baselines, and improves by more than 13% and 7% of accuracy on text classification and question answering tasks, respectively.