Why do dialogue systems need probabilistic reasoning?
Explores whether deterministic flowchart-based dialogue systems can handle realistic speech recognition error rates of 15-30 percent, and what alternative approaches might be necessary.
Can models learn to abstain when uncertain about predictions? Can dialogue planning balance fast responses with strategic depth? Can skipping transcription make voice assistants faster? Why do language models fail in gradually revealed conversations? Why do AI conversations reliably break down after multiple turns?
Calibrated models outperform larger ones by abstaining on uncertain predictions Dual-process dialogue planning balances efficiency with strategic depth through uncertainty switching Skipping transcription reduces voice assistant latency to 226 milliseconds Language models make unrecoverable premature assumptions in underspecified conversation. Multi-turn conversation failures stem from intent misalignment, not capability limits
What moves become possible when you represent ASR as a noisy observation model? How do belief distributions help systems recover from speech recognition errors? Does the same uncertainty-driven logic appear in other conversation systems? Can dialogue systems abstain from responding when uncertainty is too high? Why does dialogue-shaped text fail to produce dialogue-like operations in practice? Can systems guide users adaptively without imposing predetermined dialogue structures? Can visual representation of dialogue reveal patterns that numbers and statistics cannot? What speaker selection protocol prevents both stalling and premature convergence? Can offline reinforcement learning improve dialogue policy baseline performance? What paired speech data is needed to train end-to-end models? How do probabilistic dialogue systems handle ASR errors differently? What skills do users need to work effectively with stochastic outputs?See all 28 inquiring lines on this note →