During a conversation, there can come certain moments where its outcome hangs in the balance. In these pivotal moments, how one responds can put the conversation on substantially different trajectories leading to significantly different outcomes. Systems that can detect when such moments arise could assist conversationalists in domains with highly consequential outcomes, such as mental health crisis counseling. In this work, we introduce an unsupervised computational method for detecting such pivotal moments as they happen, in an online fashion. Our approach relies on the intuition that a moment is pivotal if our expectation of the outcome varies widely depending on what might be said next. By applying our method to crisis counseling conversations, we first validate it by showing that it aligns with human perception -- counselors take significantly longer to respond during moments detected by our method -- and with the eventual conversational trajectory -- which is more likely to change course at these times. We then use our framework to explore the relation of the counselor's response during pivotal moments with the eventual outcome of the session.

You've probably heard plenty about companies using text analytics to track of their brand on social media, or pull insights from product reviews, but the same software can predict far more serious occurrences than a dissatisfied customer. We've invited Karthik Dinakar, PhD candidate and Reid Hoffman Fellow at MIT, to speak at our Human Language Technology Conference (HLTCon) March 31st to discuss how text analytics can save lives. Karthik will present his research that applies text understanding to five areas--online cyberbullying; online adolescent distress; crisis counseling on text hotlines; ways to model, predict and treat self-harm; and "cardiolinguistics" to model atypical angina in coronary heart disease. For example, natural language processing technology can "read" texts written in social media or on chat helplines and find signals that alert doctors or counselors to a person who is threatened by cyberbullying or in imminent danger of directed self-harm. Similarly, heart attacks in women are often undiagnosed because their description of symptoms may not "present" as a heart attack.