BOLT: Fast Energy-based Controlled Text Generation with Tunable Biases
Xin Liu, Muhammad Khalifa, Lu Wang · 2023
Energy-based models (EBMs) have gained popularity for controlled text generation due to their high applicability to a wide range of constraints.However, sampling from EBMs is non-trivial, as it often requires a large number of iterations to converge to plausible text, which slows down the decoding process and makes it less practical for real-world applications.In this work, we propose BOLT, which relies on tunable biases to directly adjust the language model's output logits.Unlike prior work, BOLT maintains the generator's autoregressive nature to assert a strong control on token-wise conditional dependencies and overall fluency, and thus converges faster.When compared with state-of-the-arts on controlled generation tasks using both soft constraints (e.g., sentiment control) and hard constraints (e.g., keyword-guided topic control), BOLT demonstrates significantly improved efficiency and fluency.On sentiment control, BOLT is 7x faster than competitive baselines, and more fluent in 74.4% of the evaluation samples according to human judges.