Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit Profiles
26 Feb 2025 · Kuang Wang, Xianfei Li, Shenghao Yang, Li Zhou, Feng Jiang, Haizhou Li ·
User simulators are crucial for replicating human interactions with dialogue systems, supporting both collaborative training and automatic evaluation, especially for large language models (LLMs). However, existing simulators often rely solely on text utterances, missing implicit user traits such as personality, speaking style, and goals. In contrast, persona-based methods lack generalizability, as they depend on predefined profiles of famous individuals or archetypes. To address these challenges, we propose User Simulator with implicit Profiles (#USP), a framework that infers implicit user profiles from human-machine conversations and uses them to generate more personalized and realistic dialogues. We first develop an LLM-driven extractor with a comprehensive profile schema. Then, we refine the simulation through conditional supervised fine-tuning and reinforcement learning with cycle consistency, optimizing it at both the utterance and conversation levels. Finally, we adopt a diverse profile sampler to capture the distribution of real-world user profiles. Experimental results demonstrate that USP outperforms strong baselines in terms of authenticity and diversity while achieving comparable performance in consistency. Furthermore, dynamic multi-turn evaluations based on USP strongly align with mainstream benchmarks, demonstrating its effectiveness in real-world applications . Paper: https://arxiv.org/pdf/2502.18968v1.pdf
Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit Profiles
26 Feb 2025 · Kuang Wang, Xianfei Li, Shenghao Yang, Li Zhou, Feng Jiang, Haizhou Li ·
User simulators are crucial for replicating human interactions with dialogue systems, supporting both collaborative training and automatic evaluation, especially for large language models (LLMs). However, existing simulators often rely solely on text utterances, missing implicit user traits such as personality, speaking style, and goals. In contrast, persona-based methods lack generalizability, as they depend on predefined profiles of famous individuals or archetypes. To address these challenges, we propose User Simulator with implicit Profiles (#USP), a framework that infers implicit user profiles from human-machine conversations and uses them to generate more personalized and realistic dialogues. We first develop an LLM-driven extractor with a comprehensive profile schema. Then, we refine the simulation through conditional supervised fine-tuning and reinforcement learning with cycle consistency, optimizing it at both the utterance and conversation levels. Finally, we adopt a diverse profile sampler to capture the distribution of real-world user profiles. Experimental results demonstrate that USP outperforms strong baselines in terms of authenticity and diversity while achieving comparable performance in consistency. Furthermore, dynamic multi-turn evaluations based on USP strongly align with mainstream benchmarks, demonstrating its effectiveness in real-world applications . Paper: https://arxiv.org/pdf/2502.18968v1.pdf
Telegram and Signal Havens for Right-Wing Extremists
Since the violent storming of Capitol Hill and subsequent ban of former U.S. President Donald Trump from Facebook and Twitter, the removal of Parler from Amazon’s servers, and the de-platforming of incendiary right-wing content, messaging services Telegram and Signal have seen a deluge of new users. In January alone, Telegram reported 90 million new accounts. Its founder, Pavel Durov, described this as “the largest digital migration in human history.” Signal reportedly doubled its user base to 40 million people and became the most downloaded app in 70 countries. The two services rely on encryption to protect the privacy of user communication, which has made them popular with protesters seeking to conceal their identities against repressive governments in places like Belarus, Hong Kong, and Iran. But the same encryption technology has also made them a favored communication tool for criminals and terrorist groups, including al Qaeda and the Islamic State.
Pinterest (PINS) Stock Sinks As Market Gains
Pinterest (PINS) closed at $71.75 in the latest trading session, marking a -0.18% move from the prior day. This change lagged the S&P 500's daily gain of 0.1%. Meanwhile, the Dow gained 0.9%, and the Nasdaq, a tech-heavy index, lost 0.59%.
Heading into today, shares of the digital pinboard and shopping tool company had lost 17.41% over the past month, lagging the Computer and Technology sector's loss of 5.38% and the S&P 500's gain of 0.71% in that time.
Investors will be hoping for strength from PINS as it approaches its next earnings release. The company is expected to report EPS of $0.07, up 170% from the prior-year quarter. Our most recent consensus estimate is calling for quarterly revenue of $467.87 million, up 72.05% from the year-ago period.
