The rapid proliferation of speech-enabled technologies, including virtual assistants, video conferencing platforms, and wearable devices, has raised significant privacy concerns, particularly regarding the inference of sensitive emotional information from audio data. Existing privacy-preserving methods often compromise usability and security, limiting their adoption in practical scenarios. This paper introduces a novel, user-centric approach that leverages familiar audio editing techniques, specifically pitch and tempo manipulation, to protect emotional privacy without sacrificing usability. By analyzing popular audio editing applications on Android and iOS platforms, we identified these features as both widely available and usable. We rigorously evaluated their effectiveness against a threat model, considering adversarial attacks from diverse sources, including Deep Neural Networks (DNNs), Large Language Models (LLMs), and and reversibility testing. Our experiments, conducted on three distinct datasets, demonstrate that pitch and tempo manipulation effectively obfuscates emotional data. Additionally, we explore the design principles for lightweight, on-device implementation to ensure broad applicability across various devices and platforms.

The pervasiveness of proprietary language models has raised critical privacy concerns, necessitating advancements in private inference (PI), where computations are performed directly on encrypted data without revealing users' sensitive information. While PI offers a promising solution, its practical deployment is hindered by substantial communication and latency overheads, primarily stemming from nonlinear operations. To address this, we introduce an information-theoretic framework to characterize the role of nonlinearities in decoder-only language models, laying a principled foundation for optimizing transformer-architectures tailored to the demands of PI. By leveraging Shannon's entropy as a quantitative measure, we uncover the previously unexplored dual significance of nonlinearities: beyond ensuring training stability, they are crucial for maintaining attention head diversity. Specifically, we find that their removal triggers two critical failure modes: {\em entropy collapse} in deeper layers that destabilizes training, and {\em entropic overload} in earlier layers that leads to under-utilization of Multi-Head Attention's (MHA) representational capacity. We propose an entropy-guided attention mechanism paired with a novel entropy regularization technique to mitigate entropic overload. Additionally, we explore PI-friendly alternatives to layer normalization for preventing entropy collapse and stabilizing the training of LLMs with reduced-nonlinearities. Our study bridges the gap between information theory and architectural design, establishing entropy dynamics as a principled guide for developing efficient PI architectures. The code and implementation are available at https://github.com/Nandan91/entropy-guided-attention-llm

Fine-tuning large language models (LLMs) for specific tasks introduces privacy risks, as models may inadvertently memorise and leak sensitive training data. While Differential Privacy (DP) offers a solution to mitigate these risks, it introduces significant computational and performance trade-offs, particularly with standard fine-tuning approaches. Previous work has primarily focused on full-parameter updates, which are computationally intensive and may not fully leverage DPs potential in large models. In this work, we address these shortcomings by investigating Parameter-Efficient Fine-Tuning (PEFT) methods under DP constraints. We show that PEFT methods achieve comparable performance to standard fine-tuning while requiring fewer parameters and significantly reducing privacy leakage. Furthermore, we incorporate a data poisoning experiment involving intentional mislabelling to assess model memorisation and directly measure privacy risks. Our findings indicate that PEFT methods not only provide a promising alternative but also serve as a complementary approach for privacy-preserving, resource-efficient fine-tuning of LLMs.

The workshop 'AI-based Planning for Cyber-Physical Systems', which took place on February 26, 2024, as part of the 38th Annual AAAI Conference on Artificial Intelligence in Vancouver, Canada, brought together researchers to discuss recent advances in AI planning methods for Cyber-Physical Systems (CPS). CPS pose a major challenge due to their complexity and data-intensive nature, which often exceeds the capabilities of traditional planning algorithms. The workshop highlighted new approaches such as neuro-symbolic architectures, large language models (LLMs), deep reinforcement learning and advances in symbolic planning. These techniques are promising when it comes to managing the complexity of CPS and have potential for real-world applications.

The availability of massive amounts of data, coupled with high-performance cloud computing platforms, has driven significant progress in artificial intelligence and, in particular, machine learning and optimization. It has profoundly impacted several areas, including computer vision, natural language processing, and transportation. However, the use of rich data sets also raises significant privacy concerns: They often reveal personal sensitive information that can be exploited, without the knowledge and/or consent of the involved individuals, for various purposes including monitoring, discrimination, and illegal activities. The second AAAI Workshop on Privacy-Preserving Artificial Intelligence (PPAI-21) held at the Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21) builds on the success of last year's AAAI PPAI to provide a platform for researchers, AI practitioners, and policymakers to discuss technical and societal issues and present solutions related to privacy in AI applications. The workshop will focus on both the theoretical and practical challenges related to the design of privacy-preserving AI systems and algorithms and will have strong multidisciplinary components, including soliciting contributions about policy, legal issues, and societal impact of privacy in AI. Finally, the workshop will welcome papers that describe the release of privacy-preserving benchmarks and data sets that can be used by the community to solve fundamental problems of interest, including in machine learning and optimization for health systems and urban networks, to mention but a few examples.

Stats Perform, the revolutionary leader in sports AI and data, announced that Chief Scientist Dr. Patrick Lucey will deliver keynote remarks at the Association for the Advancement of Artificial Intelligence (AAAI-20) Workshop in Team Sport in New York on Saturday, February 8. Dr. Lucey's presentation "Interactive Sports Analytics" will examine new ways to break down player or team performance using big data and AI software. The presentation will include examples of how coaches can draw up and search for specific plays and, using AI and Stats Perform's decades of tracking and multi-agent trajectory data, simulate likely outcomes specific to a particular opponent and the players involved. In addition, Dr. Lucey will demonstrate the capabilities of new body-pose data made possible through Stats Perform's state of the art AutoSTATS technology. "We have reached an exciting moment in sports where coaches and analysts can now leverage big data and AI to generate advanced insights on play development and likely outcomes," Dr. Lucey said. "Imagine a coach drawing up an X's and O's play, the same way he would on chalkboard, on an iPad and simulating likely outcomes based on different sets of offensive and defensive opponents in-play. Imagine then being able to search that play and find video of every time a near similar play was run. With AI and big data, we are already making that happen at Stats Perform and I can't wait to meet and discuss this with the illustrious group of researchers at the AAAI Workshop."

Default and auto-epistemic reasoning were also well represented, with a number of papers discussing aspects, applications, and implementations of default reasoning systerns. Several papers emphasized nonmonotonic facets of computational vision, natural language understanding, and conimo1i-sense reasoning. Thursday evening, a panel discussion was held, with John McCarthy, Dana Scott, and Richmond Thomason as panelists. Compare it with a merely COMMON LISP (Golden Common Lisp@ Version 1.OO): Golden Common Lisp is a registered trademark of Gold Hill Computers. Our low-key, dignified approach to matchingquality candidates with quality companies will offer you the opportunity to examine your alternatives in a confidential, systematic fashion Openingsarenationwide.

Research on constraints and agents is emerging at the intersection of the communities studying constraint computation and software agents. Constraint- based reasoning systems can be enhanced by using agents with multiple problem-solving approaches or diverse problem representations. The constraint computation paradigm can be used to model agent consultation, cooperation, and competition. An interesting theme in agent interaction, which is studied here in constraint-based terms, is confronting ignorance: the agent's own ignorance or its ignorance of other agents.

Kambhampati presented theoretical planning systems is difficult. Although national AI conference, was lively It was noted that comparing planners encoding expert knowledge is at the and interesting. Both the theoretical is similar in difficulty to comparing heart of HTN planning, there and practical sides of the AI planning programming languages (in fact, the remains a considerable gap to bridge community were represented, input specifications to a planner can in using expert planning knowledge and both sides seemed to understand be viewed as a programming language). Shlomo Zilberstein (University Third, it was generally acknowledged Several papers contributed further of Massachusetts) presented a that common plan representations to the theoretical analysis of number of evaluation measures. A algorithms or through empirical An integrated system that executes common representation would allow studies (Christer Backstrom, or uses the generated plans formal comparisons among widely Linkoping University, Sweden; Subbarao should be evaluated instead of simply different planning technologies.

We summarize the Among the workshop's principal The in using these systems, and (6) computer represent the same knowledge differently workshop on cooperation among environments that facilitate to optimize their particular use heterogeneous intelligent agents, cooperation among human problem of it, or agents could obtain knowledge held July 15 during the 1991 National solvers of diverse abilities. DAI system can use as agents a collection and Edmund Durfee. It was designed Fifty submissions were received, and of existing knowledge-based to bring together researchers and 43 contributors were invited to the systems that have been developed practitioners who are studying how workshop. The workshop had four under a variety of implementation to enable a heterogeneous collection sessions that covered the topics of philosophies. In particular, representations create a special type of agent that is Fifth, agents negotiate and converge must be agreed on able to act as a broker to each of the on decisions by making deals (either before invocation or as a existing agents that need to participate under various types of pressure. Methods must also in a blackboard architecture, so it can be created for agents to assimilate cooperate with other agents.