Generating diverse and factual text is challenging and is receiving increasing attention. By sampling from the latent space, variational autoencoder-based models have recently enhanced the diversity of generated text. However, existing research predominantly depends on summarization models to offer paragraph-level semantic information for enhancing factual correctness. The challenge lies in effectively generating factual text using sentence-level variational autoencoder-based models. In this paper, a novel model called fact-aware conditional variational autoencoder is proposed to balance the factual correctness and diversity of generated text. Specifically, our model encodes the input sentences and uses them as facts to build a conditional variational autoencoder network. By training a conditional variational autoencoder network, the model is enabled to generate text based on input facts. Building upon this foundation, the input text is passed to the discriminator along with the generated text. By employing adversarial training, the model is encouraged to generate text that is indistinguishable to the discriminator, thereby enhancing the quality of the generated text. To further improve the factual correctness, inspired by the natural language inference system, the entailment recognition task is introduced to be trained together with the discriminator via multi-task learning. Moreover, based on the entailment recognition results, a penalty term is further proposed to reconstruct the loss of our model, forcing the generator to generate text consistent with the facts. Experimental results demonstrate that compared with competitive models, our model has achieved substantial improvements in both the quality and factual correctness of the text, despite only sacrificing a small amount of diversity. Furthermore, when considering a comprehensive evaluation of diversity and quality metrics, our model has also demonstrated the best performance.

Memory leak is a common software vulnerability that can decrease the reliability of an application and, in severe cases, even cause program crashes. If there are intentionally triggerable memory leak vulnerabilities in a program, attackers can exploit these bugs to launch denial-of-service attacks or induce the program to exhibit unexpected behaviors due to low memory conditions. Existing fuzzing techniques primarily focus on improving code coverage, and specialized fuzzing techniques for individual memory-related defects like uncontrolled memory allocation do not address memory leak vulnerabilities. MemLock is the first fuzzing technique to address memory consumption vulnerabilities including memory leakage. However, the coverage-centric guidance mechanism of MemLock introduces a degree of aimlessness in the testing process, that results in low seed quality and slow bug exposure speed. To address this issue, we propose a risk areas guidance-based fuzzing technique called RBZZER. First, RBZZER retains MemLock’s memory consumption-guided mechanism and introduces a novel distance-guided approach to expedite the arrival of fuzzing at the potential memory areas. Second, we introduce a new seed scheduling strategy called risk areas-based seed scheduling, which classifies seeds based on potential memory leak areas in the program and further schedules them, thereby effectively improving the efficiency of discovering memory leak vulnerabilities. Experiments demonstrate that RBZZER outperforms the state-of-the-art fuzzing techniques by finding 52% more program unique crashes than the second-best counterpart. In particular, RBZZER can discover the amount of memory leakage at least 112% more than the other baseline fuzzers. Besides, RBZZER detects memory leaks at an average speed that is 9.10x faster than MemLock.