Pulsar is a software platform for social media monitoring, audience intelligence and social listening that allows organizations to monitor and analyze online conversations across social media, news, and other digital sources. The platform combines social media listening, media monitoring, trend analysis, and audience segmentation to help users understand public discussions and audience behavior in real time. The platform is a social listening platform, which aggregates data from networks such as X, Facebook, Instagram, and forums) and applies artificial intelligence for text and sentiment analysis. Pulsar is offered as a cloud-based Software as a Service (SaaS) tool and insights consultancy. It has been part of Pulsar Group (formerly Access Intelligence), a publicly listed group of communications software products, since 2019. As well as commercial uses, the platform has been used in peer-reviewed academic research analysing online discourse. The platform is listed on the UK government's G-Cloud 14 Digital Marketplace for the provision of social listening and audience intelligence services. == History == Pulsar originated in the early 2010s as a project within Face, a London-based innovation and market research consultancy. The platform's first product, Pulsar TRAC, launched in 2013 as a social media analytics tool. Pulsar TRAC was designed to measure the reach of conversations, mapping brand audiences, and tracking how content spreads through networks. The development was led by Dr Francesco D'Orazio, who created the Pulsar brand and led the development of the platform while serving as VP of Product and Innovation at Face. Face itself had been acquired by the Cello Group Plc (a UK-based advisory firm) in 2012, and Pulsar became part of Cello's portfolio of research and data tools. In January 2017, Cello Group made a significant investment to scale Pulsar and announced the merger of Face's qualitative research business into Pulsar, unifying both under the Pulsar brand for global expansion. In 2018, Pulsar opened an office in Los Angeles to better serve its growing U.S. client base in media, healthcare, and entertainment sectors and Francesco D'Orazio was appointed CEO. The company focused on developing new products amid a wave of consolidation in the social listening industry. In October 2019, Pulsar was acquired by Access Intelligence Plc (now Pulsar Group), an AIM-listed communications software company. The group, which also owns PR and media tools Isentia, Vuelio and ResponseSource, integrated Pulsar to their end-to-end marketing and communications insights offering. Pulsar established a new office in Sydney, Australia in 2022 as part of this global expansion, adding to its existing offices in London and Los Angeles. In 2023, Pulsar Group (then Access Intelligence) was recognised as one of Europe's fastest growing companies by the Financial Times. In May 2024, Access Intelligence PLC changed its name to Pulsar Group PLC. The company has since continued to develop its platform. In March 2025 it introduced new tool Narratives AI, described as a "search engine for public opinion" and the first of its kind for analyzing public narratives and their evolutions in both social media and the news. In October 2025, Pulsar launched Insight Agents, a set of AI agents embedded into the platform advertised to "proactively anticipate user needs or issues, carry out routine tasks, uncover anomalies in your datasets, and prompt responses at scale, 24/7." == Products == Pulsar's architecture integrates four main products into a single interface. The core product suite is often broken into three main components: Pulsar TRAC (for social listening and audience analysis), Pulsar TRENDS (for trend discovery and analysis), and Pulsar CORE (for owned-channel and web analytics). Pulsar's fourth product is Narratives AI. === Pulsar TRAC === Pulsar TRAC is a social listening and audience intelligence platform that allows users to configure searches that track public conversations and measure audience behaviour. Pulsar TRAC is focused on conversation insights and audience segmentations - the platform is reported to collect and analyse data from a wide range of sources, including major social networks, forums, news and review sites, and ecommerce platforms, with real-time visualisations and AI-supported analytics used to find patterns and communities of interest. Pulsar TRAC can be incorporated into workflows with other audience tools, such as an integration with Audiense that connects TRAC's conversation insights to external audience-segmentation datasets. === Pulsar CORE === Pulsar CORE centres on the analysis of owned-channel data, such as brand social media profiles, website interaction and other in-house digital assets, to generate audience and content insights. CORE can monitor published content, evaluate competitors, and extract demographic and behavioural segmentation from owned channels. === Narratives AI === Narratives AI is a tool within the Pulsar audience intelligence platform that uses artificial intelligence to detect, cluster and analyse narratives forming across social and news media. It was launched in March 2025 as a standalone search interface that processes real-time and historical data to find cultural trends, behaviours and beliefs. It uses clustering algorithms and visualisation to show how conversations form and spread online, and their relative importance within wider discourse. == Notable features == === Insight Agents === Pulsar's Insight Agents are AI-powered agents within the Pulsar platform designed to automate and augment common tasks in media, social, audience and narrative intelligence. Branded as TeamMates, these agents are grouped into four functional types: Sentinels for real-time monitoring, anomaly detection and alerting Oracles for forecasting and scenario planning Custodians for governance, compliance and policy enforcement Analysts for research, reporting and recommendations Each agent is trained on Pulsar's multi-source data and domain-specific workflows. In February 2026, Pulsar introduced 'Crisis Oracle,' an AI-driven system designed to quantify narrative momentum and predict reputational risk. == Academic research == Pulsar has been used as a data collection and analysis tool in peer-reviewed academic research across public health, infodemiology, veterinary science, and policy research. Published uses include a World Health Organization report on infodemic management, a Journal of Medical Internet Research study on headache and migraine discourse across Japan, Germany, and France, a Frontiers in Big Data study of Long COVID narratives, and Frontiers in Veterinary Science studies on canine chronic kidney disease and oral medication administration in dogs.
Automated negotiation
Automated negotiation is a form of interaction in systems that are composed of multiple autonomous agents, in which the aim is to reach agreements through an iterative process of making offers. Automated negotiation can be employed for many tasks human negotiators regularly engage in, such as bargaining and joint decision making. The main topics in automated negotiation revolve around the design of protocols and negotiating strategies. == History == Through digitization, the beginning of the 21st century has seen a growing interest in the automation of negotiation and e-negotiation systems, for example in the setting of e-commerce. This interest is fueled by the promise of automated agents being able to negotiate on behalf of human negotiators, and to find better outcomes than human negotiators. == Examples == Examples of automated negotiation include: Online dispute resolution, in which disagreements between parties are settled. Sponsored search auction, where bids are placed on advertisement keywords. Content negotiation, in which user agents negotiate over HTTP about how to best represent a web resource. Negotiation support systems, in which negotiation decision-making activities are supported by an information system.
Top 10 AI Customer-support Bots Compared (2026)
Trying to pick the best AI customer-support bot? An AI customer-support bot is software that uses machine learning to help you get more done — it scales effortlessly from a single task to thousands. The best picks balance beginner-friendly simplicity with the depth power users need, and they ship updates often. Whether you are a beginner or a pro, the right AI customer-support bot slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
Emergent (software)
Emergent (formerly PDP++) is a biologically-based neural simulation software that is primarily intended for creating models of the brain and cognitive processes. Development initially began in 1995 at Carnegie Mellon University, and as of 2014, continues at the University of Colorado at Boulder. The 3.x release of the software, which was known as PDP++, is featured in the textbook Computational Explorations in Cognitive Neuroscience. == Features == Emergent features a modular design, based on the principles of object-oriented programming. It runs on Microsoft Windows, Darwin / macOS and Linux. C-Super-Script (variously, CSS and C^C), a built-in C++-like interpreted scripting language, allows access to virtually all simulator objects and can initiate all the same actions as the GUI, and more. Version 4 and upward features a full 3D environment for visualizations, based on Qt and Open Inventor. Robotics simulations are made possible by integration with the Open Dynamics Engine. A plugin system allows for expanding the software in many ways. Version 5 introduced parallel threading support, numerous speed improvements, a help browser featuring an interface to the project's Wiki and auto-generated documentation, undo and redo using diffs and a definable undo depth. In addition, 5.0.2 introduced a built-in plugin source code editor, and plugins can now be compiled from the main interface, enabling full development of plugins within Emergent. Emergent also provides an implementation of Leabra which was developed by Randall C. O'Reilly in his PhD thesis.
Deborah Raji
Inioluwa Deborah Raji (born 1995/1996) is a Nigerian-Canadian computer scientist and socio-tech leader who works on algorithmic bias, AI accountability, and algorithmic auditing. A current Mozilla fellow, she has been recognized by MIT Technology Review and Forbes as one of the world's top young innovators. Raji started her work with racial bias in technology during her internship with Clarifai when she recognized that people of color were more often tagged for NSFW compared to white people. Raji has previously worked with Joy Buolamwini, Timnit Gebru, and the Algorithmic Justice League on researching gender and racial bias in facial recognition technology. Her work on racial bias in facial recognition has forced companies to ultimately change their practices. She has also worked with Google’s Ethical AI team and been a research fellow at the Partnership on AI and AI Now Institute at New York University working on how to operationalize ethical considerations in machine learning engineering practice. She was working on a computer vision model that would help clients flag inappropriate images as NSFW. == Early life and education == Raji was born in Port Harcourt, Nigeria, and moved to Mississauga, Ontario, Canada, when she was four years old. Eventually her family moved to Ottawa. She attended Colonel By Secondary School and completed the International Baccalaureate programme. She studied Engineering Science at the University of Toronto, graduating in 2019. In 2015, she founded Project Include, a nonprofit providing increased student access to engineering education, mentorship, and resources in low income and immigrant communities in the Greater Toronto Area. She started a Doctor of Philosophy - PhD, in Computer Science from the University of California, Berkeley in Aug 2021. == Career and research == Raji worked with Joy Buolamwini at the MIT Media Lab and Algorithmic Justice League, where she audited commercial facial recognition technologies from Microsoft, Amazon, IBM, Face++, and Kairos. They found that these technologies were significantly less accurate for darker-skinned women than for white men. With support from other top AI researchers and increased public pressure and campaigning, their work led IBM and Amazon to agree to support facial recognition regulation and later halt the sale of their product to police for at least a year. Raji also interned at machine learning startup Clarifai, where she worked on a computer vision model for flagging images. She participated in a research mentorship program at Google and worked with their Ethical AI team on creating model cards, a documentation framework for more transparent machine learning model reporting. She also co-led the development of internal auditing practices at Google. Her contributions at Google were separately presented and published at the AAAI conference and ACM Conference on Fairness, Accountability, and Transparency. In 2019, Raji was a summer research fellow at The Partnership on AI working on setting industry machine learning transparency standards and benchmarking norms. Raji was a Tech Fellow at the AI Now Institute worked on algorithmic and AI auditing. Currently, she is a fellow at the Mozilla Foundation researching algorithmic auditing and evaluation. Raji's work on bias in facial recognition systems has been highlighted in the 2020 documentary Coded Bias directed by Shalini Kantayya. She also took part in the 2026 documentary The AI Doc: Or How I Became an Apocaloptimist directed by Daniel Roher. == Awards == 2019 Venture Beat AI Innovations Award in category AI for Good (received with Joy Buolamwini and Timnit Gebru) 2020 MIT Technology Review 35 Under 35 Innovator Award 2020 EFF Pioneer Award (received with Buolamwini and Gebru) 2021 Forbes 30 Under 30 Award in Enterprise Technology 2021 100 Brilliant Women in AI Ethics Hall of Fame Honoree 2023 Time magazine 100 Most Influential People in AI
Model compression
Model compression is a machine learning technique for reducing the size of trained models. Large models can achieve high accuracy, but often at the cost of significant resource requirements. Compression techniques aim to compress models without significant performance reduction. Smaller models require less storage space, and consume less memory and compute during inference. Compressed models enable deployment on resource-constrained devices such as smartphones, embedded systems, edge computing devices, and consumer electronics computers. Efficient inference is also valuable for large corporations that serve large model inference over an API, allowing them to reduce computational costs and improve response times for users. Model compression is not to be confused with knowledge distillation, in which a smaller "student" model is trained to imitate the input-output behavior of a larger "teacher" model (as opposed to using the "teacher"'s trained parameters or the "teacher"'s training targets). == Techniques == Several techniques are employed for model compression. === Pruning === Pruning sparsifies a large model by setting some parameters to exactly zero. This effectively reduces the number of parameters. This allows the use of sparse matrix operations, which are faster than dense matrix operations. Pruning criteria can be based on magnitudes of parameters, the statistical pattern of neural activations, Hessian values, etc. === Quantization === Quantization reduces the numerical precision of weights and activations. For example, instead of storing weights as 32-bit floating-point numbers, they can be represented using 8-bit integers. Low-precision parameters take up less space, and takes less compute to perform arithmetic with. It is also possible to quantize some parameters more aggressively than others, so for example, a less important parameter can have 8-bit precision while another, more important parameter, can have 16-bit precision. Inference with such models requires mixed-precision arithmetic. Quantized models can also be used during training (rather than after training). PyTorch implements automatic mixed-precision (AMP), which performs autocasting, gradient scaling, and loss scaling. === Low-rank factorization === Weight matrices can be approximated by low-rank matrices. Let W {\displaystyle W} be a weight matrix of shape m × n {\displaystyle m\times n} . A low-rank approximation is W ≈ U V T {\displaystyle W\approx UV^{T}} , where U {\displaystyle U} and V {\displaystyle V} are matrices of shapes m × k , n × k {\displaystyle m\times k,n\times k} . When k {\displaystyle k} is small, this both reduces the number of parameters needed to represent W {\displaystyle W} approximately, and accelerates matrix multiplication by W {\displaystyle W} . Low-rank approximations can be found by singular value decomposition (SVD). The choice of rank for each weight matrix is a hyperparameter, and jointly optimized as a mixed discrete-continuous optimization problem. The rank of weight matrices may also be pruned after training, taking into account the effect of activation functions like ReLU on the implicit rank of the weight matrices. == Training == Model compression may be decoupled from training, that is, a model is first trained without regard for how it might be compressed, then it is compressed. However, it may also be combined with training. The "train big, then compress" method trains a large model for a small number of training steps (less than it would be if it were trained to convergence), then heavily compress the model. It is found that at the same compute budget, this method results in a better model than lightly compressed, small models. In Deep Compression, the compression has three steps. First loop (pruning): prune all weights lower than a threshold, then finetune the network, then prune again, etc. Second loop (quantization): cluster weights, then enforce weight sharing among all weights in each cluster, then finetune the network, then cluster again, etc. Third step: Use Huffman coding to losslessly compress the model. The SqueezeNet paper reported that Deep Compression achieved a compression ratio of 35 on AlexNet, and a ratio of ~10 on SqueezeNets.
Dialogflow
Dialogflow is a natural language understanding platform used to design and integrate a conversational user interface into mobile apps, web applications, devices, bots, interactive voice response systems and related uses. == History == In May 2012, Speaktoit received a venture round (funding terms undisclosed) from Intel Capital. In July 2014, Speaktoit closed their Series B funding led by Motorola Solutions Venture Capital with participation from new investor Plug and Play Ventures and existing backers Intel Capital and Alpine Technology Fund. In September 2014, Speaktoit released api.ai (the voice-enabling engine that powers Assistant) to third-party developers, allowing the addition of voice interfaces to apps based on Android, iOS, HTML5, and Cordova. The SDK's contain voice recognition, natural language understanding, and text-to-speech. api.ai offers a web interface to build and test conversation scenarios. The platform is based on the natural language processing engine built by Speaktoit for its Assistant application. Api.ai allows Internet of Things developers to include natural language voice interfaces in their products. Assistant and Speaktoit's websites now redirect to api.ai's website Archived 2017-10-10 at the Wayback Machine, which redirects to the Dialogflow website. Google bought the company in September 2016 and was initially known as API.AI; it provides tools to developers building apps ("Actions") for the Google Assistant virtual assistant. The organization discontinued the Assistant app on December 15, 2016. In October 2017, it was renamed as Dialogflow. In November 2017, Dialogflow became part of Google Cloud Platform.