Engagement Rewarded Actor-Critic with CQL for Speech-Driven Laughter Backchannel Generation
We propose a speech-driven laughter backchannel generation model to reward engagement during human-agent interaction. We formulate the problem as a Markov decision process where speech signal represents the state and the objective is to maximize human engagement. Since online training is often impractical in the case of human-agent interaction, we utilize the existing human-to-human dyadic interaction datasets to train our agent for the backchannel generation task. We address the problem using an actor-critic method based on conservative Q-learning (CQL), that mitigates the distributional shift problem by suppressing Q-value over-estimation during training. The proposed CQL based approach is evaluated objectively on the IEMOCAP dataset for laughter generation task. When compared to the existing off-policy Q-learning methods, we observe an improved compliance with the dataset in terms of laugh generation rate. Furthermore, we show the effectiveness of the learned policy by estimating the expected engagement us
2 views
24
5
3 days ago 00:05:53 1
Binance Launchpool | Launchpool Binance | Binance Launchpool for Beginners
3 days ago 00:07:35 1
Crypto Games | NFT Game | Crypto Gaming
3 days ago 00:02:58 1
Unlocked with Lando Norris & Oscar Piastri - Episode 3: Post-Race Nutrition
4 days ago 00:04:53 1
Crypto Presale New🔥 BlockDAG: 🔥The Ultimate Breakdown of the Top ICO & Presale
5 days ago 00:32:05 1
EXTREMELY BAD NEWS FOR BITCOIN BEARS !!!
6 days ago 00:29:22 2
SILENT HILL 2 REMAKE VR is Mind Blowing! // Face to Face With PYRAMID HEAD (UEVR)