Photoplethysmography (PPG)-based foundation models are gaining traction due to the widespread use of PPG in biosignal monitoring and their potential to generalize across diverse health applications. In this paper, we introduce Pulse-PPG, the first open-source PPG foundation model trained exclusively on raw PPG data collected over a 100-day field study with 120 participants. Existing PPG foundation models are either open-source but trained on clinical data or closed-source, limiting their applicability in real-world settings. We evaluate Pulse-PPG across multiple datasets and downstream tasks, comparing its performance against a state-of-the-art foundation model trained on clinical data. Our results demonstrate that Pulse-PPG, trained on uncurated field data, exhibits superior generalization across clinical and mobile health applications in both lab and field settings. This suggests that exposure to real-world variability enables the model to learn fine-grained representations, making it more adaptable across tasks. Furthermore, pre-training on field data surprisingly outperforms its pre-training on clinical data in many tasks, reinforcing the importance of training on real-world, diverse datasets. To encourage further advancements in robust foundation models leveraging field data, we plan to release Pulse-PPG, providing researchers with a powerful resource for developing more generalizable PPG-based models.

There is widespread concern about the negative impacts of social media feed ranking algorithms on political polarization. Leveraging advancements in large language models (LLMs), we develop an approach to re-rank feeds in real-time to test the effects of content that is likely to polarize: expressions of antidemocratic attitudes and partisan animosity (AAPA). In a preregistered 10-day field experiment on X/Twitter with 1,256 consented participants, we increase or decrease participants' exposure to AAPA in their algorithmically curated feeds. We observe more positive outparty feelings when AAPA exposure is decreased and more negative outparty feelings when AAPA exposure is increased. Exposure to AAPA content also results in an immediate increase in negative emotions, such as sadness and anger. The interventions do not significantly impact traditional engagement metrics such as re-post and favorite rates. These findings highlight a potential pathway for developing feed algorithms that mitigate affective polarization by addressing content that undermines the shared values required for a healthy democracy.

Chatbots or conversational agents (CAs) are increasingly used to improve access to digital psychotherapy. Many current systems rely on rigid, rule-based designs, heavily dependent on expert-crafted dialogue scripts for guiding therapeutic conversations. Although recent advances in large language models (LLMs) offer the potential for more flexible interactions, their lack of controllability and transparency poses significant challenges in sensitive areas like psychotherapy. In this work, we explored how aligning LLMs with expert-crafted scripts can enhance psychotherapeutic chatbot performance. Our comparative study showed that LLMs aligned with expert-crafted scripts through prompting and fine-tuning significantly outperformed both pure LLMs and rule-based chatbots, achieving a more effective balance between dialogue flexibility and adherence to therapeutic principles. Building on findings, we proposed ``Script-Strategy Aligned Generation (SSAG)'', a flexible alignment approach that reduces reliance on fully scripted content while enhancing LLMs' therapeutic adherence and controllability. In a 10-day field study, SSAG demonstrated performance comparable to full script alignment and outperformed rule-based chatbots, empirically supporting SSAG as an efficient approach for aligning LLMs with domain expertise. Our work advances LLM applications in psychotherapy by providing a controllable, adaptable, and scalable solution for digital interventions, reducing reliance on expert effort. It also provides a collaborative framework for domain experts and developers to efficiently build expertise-aligned chatbots, broadening access to psychotherapy and behavioral interventions.

Detecting and normalizing temporal expressions is an essential step for many NLP tasks. While a variety of methods have been proposed for detection, best normalization approaches rely on hand-crafted rules. Furthermore, most of them have been designed only for English. In this paper we present a modular multilingual temporal processing system combining a fine-tuned Masked Language Model for detection, and a grammar-based normalizer. We experiment in Spanish and English and compare with HeidelTime, the state-of-the-art in multilingual temporal processing. We obtain best results in gold timex normalization, timex detection and type recognition, and competitive performance in the combined TempEval-3 relaxed value metric. A detailed error analysis shows that detecting only those timexes for which it is feasible to provide a normalization is highly beneficial in this last metric. This raises the question of which is the best strategy for timex processing, namely, leaving undetected those timexes for which is not easy to provide normalization rules or aiming for high coverage.

The media had a field day when a Google engineer recently claimed that the company's artificial intelligence technology had become "sentient." For every article joking about Skynet and HAL 9000, there was another assuming it must be true and questioning the ethics of it all. Missing in most of the coverage was any recognition of how far and fast this technology has advanced and how broadly it impacts our lives on a daily basis, in ways both large and small. It was only ten years ago on June 26, 2012 that the New York Times wrote about Google's deep learning discovery using machine learning, essentially teaching a computer to train itself with enormous amounts of data. The article was headlined How Many Computers to Identify a Cat? 16,000.

Intel is working to improve workloads for artificial intelligence (AI) developers and data scientists by offering a wide variety of AI Python optimizations and tools. Whether you are at the beginning of your AI journey and looking to learn more, or you consider yourself a data science veteran, you can find great insight about Intel's tools and technology offerings through this showcase and the accompanying resources. This Tech Field Day Showcase is presented by Todd Tomashek, Software Engineering Manager in Machine Learning Performance at Intel. Chris Grundemann is a passionate, creative technologist and a strong believer in technology's power to aid in the betterment of humankind. In his current role as Managing Director at Grundemann Technology Solutions he is expressing that passion by helping technology businesses grow and by helping any business grow with technology.

Find out more about BrainChip’s high-performance, small and ultra-low-power solutions that can be applied in a wide array of edge applications.

All the sessions from Transform 2021 are available on-demand now. The analyst community is having a field day with hype around "low code." IDC has predicted that there will be more and more low code used and that the worldwide population of low-code developers will grow with a CAGR of 40.4% from 2021 to 2025. Gartner predicted that low code will increase nearly 30% from 2020 to reach $5.8 billion in 2021. Forrester has also jumped on the low-code hype wagon and forecasted that by the end of 2021, 75% of application development will use low-code platforms.

With power and physical space at a premium at the edge, how can you apply AI models without having to worry about slowing yourself down? BrainChip offers an answer through their Akida neuromorphic chip, which they presented during their AI Field Day appearance. There's so much data at play in the modern enterprise, and for organizations that leverage Internet of Things (IoT) devices for applications like manufacturing, autonomous vehicles, and others, having a way to process and act upon the data they collect is crucial to streamlining operations. Traditionally, said data would need to be ported back to on-premises data centers so it could be consumed by the powerful processors housed there. However, to keep the pace and agility modern organizations need, passing back and forth of data must be cut down, and data needs to be processed at the edge.

Hackers are having a field day, and weak authentication is a major cause. The vast majority of cyberattacks -- some 80%, statistics show -- have their roots in compromised passwords that hackers get hold of. All it takes is one stolen password for hackers to wreak havoc; and according to experts, that single password breach can cost enterprise firms over $7 million. Many schemes have been tried to build up password security, including increased education and 2FA. But despite that, password compromise statistics remain stubbornly high, cybersecurity education programs, although widespread, don't seem to work, and 2FA has its own security issues.