Mediaboss Marketing

CAMBRIDGE, Mass.--(BUSINESS WIRE)--Dyno Therapeutics, a biotechnology company pioneering use of artificial intelligence in gene therapy, today announced a publication in the journal Science that demonstrates the power of a comprehensive machine-guided approach to engineer improved capsids for gene therapy delivery. The research was conducted by Dyno co-founders Eric D. Kelsic, Ph.D. and Sam Sinai, Ph.D., together with colleague Pierce Ogden, Ph.D., at Harvards Wyss Institute for Biologically Inspired Engineering and the Harvard Medical School laboratory of George M. Church, Ph.D., a Dyno scientific co-founder. The publication, entitled Comprehensive AAV capsid fitness landscape reveals a viral gene and enables machine-guided design, is available here.1

AAV capsids are presently the most commonly used vector for gene therapy because of their established ability to deliver genetic material to patient organs with a proven safety profile. However, there are only a few naturally occurring AAV capsids, and they are deficient in essential properties for optimal gene therapy, such as targeted delivery, evasion of the immune system, higher levels of viral production, and greater transduction efficiency. Starting at Harvard in 2015, the authors set out to overcome the limitations of current capsids by developing new machine-guided technologies to rapidly and systematically engineer a suite of new, improved capsids for widespread therapeutic use.

In the research published in Science, the authors demonstrate the advance of their unique machine-guided approach to AAV engineering. Previous approaches have been limited by the difficulty of altering a complex capsid protein without breaking its function and by the general lack of knowledge regarding how AAV capsids interact with the body. Historically, rather than addressing this challenge directly, the most popular approaches to capsid engineering have taken a roundabout solution: generating libraries of new capsids by making random changes to the protein. However, since most random changes to the capsid actually result in decreased function, such random libraries contain few viable capsids, much less improved ones. Recognizing the limitation of conventionally generated capsid libraries, the authors implemented a machine-guided approach that gathered a vast amount of data using new high-throughput measurement technologies to teach them how to build better libraries and, ultimately, lead to synthetic capsids with optimized delivery properties.

Focusing on the AAV2 capsid, the authors generated a complete landscape of all single codon substitutions, insertions and deletions, then measured the functional properties important for in vivo delivery. They then used a machine-guided approach, leveraging these data to efficiently generate diverse libraries of AAV capsids with multiple changes that targeted the mouse liver and that outperformed AAVs generated by conventional random mutagenesis approaches. In the process, the authors systematic efforts unexpectedly revealed the existence of a previously-unrecognized protein encoded within the sequence of all the most popular AAV capsids, which they termed membrane-associated accessory protein (MAAP). The authors believe that the protein plays a role in the natural life cycle of AAV.

This is just the beginning of machine-guided engineering of AAV capsids to transform gene therapy, underscores co-author Sam Sinai, Ph.D., Lead Machine Learning Scientist and co-founder of Dyno Therapeutics. The success of the simple linear models used in this study has led us to pursue more data and higher capacity machine learning models, where the potential for improvement in capsid designs feels boundless.

The results in the Science publication demonstrate, for the first time, the power of linking a comprehensive set of advanced techniques large scale DNA synthesis, pooled in vitro and in vivo screens, next-generation sequencing readouts, and iterative machine-guided capsid design to generate optimized synthetic AAV capsids, explains co-first and co-corresponding author Eric D. Kelsic, Ph.D., CEO and co-founder of Dyno Therapeutics. At Dyno, our team is committed to advancing these technologies to identify capsids that meet the urgent needs of patients who can benefit from gene therapies.

About Dyno TherapeuticsDyno Therapeutics is a pioneer in applying artificial intelligence to gene therapy. The companys powerful and proprietary genetic engineering platform is designed to rapidly and systematically develop improved AAV capsids that redefine the gene therapy landscape. Dyno was founded by experienced biotech entrepreneurs and leading scientists in the fields of synthetic biology, gene therapy, and machine learning. The company is located in Cambridge, Massachusetts. For additional information, please visit the company website at http://www.dynotx.com

See original here:

Dyno Therapeutics Announces Research Published in Science Enabling Artificial Intelligence Approach to Create New AAV Capsids for Gene Therapies -...

AI, AI Everywhere

The Artificial Intelligence field (AI) is creating a continuity that encompasses climate change science and the preparedness of the U.S. military to climate risks. This continuity appears through the central role of AI in two apparently disconnected foresight civilian and military uses.

Climate Central published in Nature a new assessment of the effects of climate change estimates. It establishes that 300 million people will be threatened by the sea-level rise and coastal flooding by 2050. In 2100, the land where 200 million people live today could be submersed daily (Climate Central, Report: Flooded Future: Global vulnerability to sea level rise worse than previously understood, October 29, 2019). This estimate is a tripling from precedent assessments. It is the result of the use of AI to correct series of datasets.

AI predicts sea-level rise and coastal flooding will threaten 300 million people by 2050.

Previously we thought 80 million people would be at risk by 2100.

During the same period, the Centre for Climate and Security published an article about a recent publication by the U.S. Army War College. The document, Implications of Climate change for the U.S Army, however, cannot be found anymore on the publications page of the U.S. Army War College. A rapid internet search allows us to find the report cited in a few articles and posted in a pdf version on internet journals, such asViceandPopular Mechanics. Yet, it cannot be found on official Department of Defense websites.

Nonetheless, this document establishes that adapting to the violent ecological, military, political, economic and social consequences of climate change is a dire and imperative necessity for the Army and for the entire U.S. military. Some parts of this report are centred on the use of artificial intelligence for force enhancement and energy use. It also calls for the modernization of training through a better and systematic use of virtual training and simulation.

In other words, artificial intelligence is creating a cognitive bridge between climate science and the U.S. military. It also creates new adaptation possibilities to the short and long term consequences of climate change.

In this article, we are going to study the strategic consequences of this scientific and military uses of AI in the climate change field. We are also going to see how the introduction of AI in both climate change and military affairs defines the emergence of a new political and planetary era.

Between now and 2100, a total of 360 (310-420)million people living on coastlines will be put at risk by flooding induced byclimate change driven sea-level rise (ClimateCentral, ibid). Compared with the currentglobal population of 7,5 billion people, it means that one person in 22 isgoing to be put at risk by this planetary trend with, at least, an annualflood, while the rise of the ocean could reach almost two metres. Those resultsare in sharp contrast with a former assessment establishing that 80 millionpeople would be at risk at the end of the century.

Now, the lowest and most densely populated coastlines, as in Bangladesh, Vietnam, China, Indonesia, Thailand, the Netherlands, and Louisiana, among others, 237 to 300 million people will be threatened by annual flooding in 2050. Those humongous numbers are the result of a new calculation. This new approach rests upon the cleaning by an AI-neural network system of the dataset previously used by scientists (Climate Central report in Nature, Scott A. Kulp and Benjamin H. Strauss, New elevation data triple estimates of global vulnerability to sea-level rise and coastal flooding, 29 October 2019).

This dataset is a compilation of the NASA and other satellite and air based lidar observations (Kulp and Strauss, ibid). The AI system corrected different results. For example, it corrected the way some space or air sensors could confuse coast altitude with city skylines altitudes. Those errors were inducing that those higher elevations were safer. So, this new neural network digital elevation model generates new results. It also generates an interactive visualization that alerts about the shape of things soon to come.

Thisstudy also establishes that, very likely, the amplitude of the sea-level risewill overwhelm the ability and resources of countries and cities to buildcoastal flood defences, as levees and seawalls. It clearly appears thatdeveloping countries as well as old industrialized countries are at risks, fromthe Vietnam to the Florida coasts.

However, the authors of the study are keen to precise that their study does not factor in several variables. Among them are the future coastal population densities, the geomorphological consequences of wetland submersion and accelerated ground erosion. The authors also precise that they have not yet integrated the socioeconomic consequences of this climate-ocean trend. Neither have they developed scenarios about the mass migrations, social unrests and conflicts that this AI-based research implies.

In a previous article, we saw how the U.S. Army research branch makes use of climate change research in order to define and propose a massive military adaptation effort (Jean-Michel Valantin, The U.S Army versus a Warming Planet, The Red (Team) Analysis Society, November 12, 2019).

In thisreport, the authors promote the use of artificial intelligence in order todevelop smart electrical and distributed grid, because The automated,A.I.-enhanced force of the Armys future is one that runs on electricity, notJP-8 (fuel). More efficient or resilient production of electricity throughmicro-nuclear power generation or improved solar arrays can fundamentally alterthe mobility and the logistical challenges of a mechanized force (p.22).

So, this recommendations aim at developing the robustness and resiliency of the U.S. Army operations in an energy constrained and climate sensitive near-future. This development will depend upon the interactions between AI and robotization. That is to say the military integration of actuators (Hlne Lavoix, Sensor and Actuator for AI: Inserting Artificial Intelligence in Reality, The Red (Team) Analysis Society, 14 January 2019). Those are the AI extension into physical reality. So, in military terms, AI will support and optimize the deployment of mechanical ground forces on theatres of operations (Hlne Lavoix, Sensor and actuator (4): Artificial Intelligence, the Long March towards Advanced Robots and Geopolitics, The Red (Team) Analysis Society, May 13, 2019).

In order to better prepare military actors to these new realities, the report also advocates for a massive use of virtual reality. Indeed, training through virtual reality simulations could help to better prepare officers and actors (Hlne Lavoix, How to Win a War with Artificial intelligence and Few Casualties, The Red (Team) Analysis Society, May 27, 2019). As it happens, they will have to handle future semi-automatized military capabilities in a world brutalized by climate change. AI would also support the responses of the U.S. military against foreign anddomestic massive cyber attacks. And it would drive the development of the U.S. military in the current technological race.

It is difficult not to think that, in the parts about the use of artificial intelligence, the authors are not alluding to the current massive militarization of AI by the Chinese military, both in training and at the operational and decision-making levels (Jean-Michel Valantin, Militarizing Artificial Intelligence China (1) and (2),The Red Team Analysis Society, April 23, 2018).

It must be kept in mind that these recommendations are part of a U.S. Army advocacy for climate change adaptation. What motivates these military recommendations is the rapid multiplication of multidimensional risks (Jean-Michel Valantin, The Midwest, the Trade war and the Swine Flu pandemic: the Agricultural and Food Super Storm is Here, The Red (Team) Analysis, June 3, 2019), as those the Climate Central report defines about sea-level rise.

As wecan see, AI becomes a central feature of the new reality landscape. As such, itbecomes a climate science tool as well as a military tool for transformationand adaptation to our warming and riskier planet.

In other terms, AI is entering the fray of the hyper siege, i.e. the cascade of consequences that are interlocking social, infrastructural, biologic vulnerabilities with climate driven events. Those cascades are becoming an entity that is besieging contemporary societies (Jean-Michel Valantin, Hyper siege: Climate Change and U.S National Security,The Red (Team) Analysis Society, March 17, 2014 and The U.S Navy vs Climate and ocean change,The Red (Team) Analysis, June 11, 2018, and David Wallace-Wells, The Unhinabitable Earth, Life After Warming, 2019).

So, AI power unveils itself (Hlne Lavoix, When Artificial Intelligence will Power Geopolitics-Presenting AI, The Red (Team) Analysis Society, November 27, 2017), through scientific research and military preparedness, as a tool and a possible ally in the face of the rapidly coming perfect climate and social super storm.

In this ecological and strategic context, AI power becomes an artificial continuum, both technological and cognitive. It actuates itself through climate research and military adaptation to the very climate change that it helps foresee. This creates an unexpected alliance between AI power, climate science and military foresight and warning. This new AI power will be useful for adapting to the planetary crisis and its cascade of hyper violent consequences (Jean-Michel Valantin, The Planetary Crisis Rules, part 1, 2, 3, 4, 5, The Red (Team) Analysis Society).

Instrategic terms, the convergence of AI power and the will and capabilities toadapt to the Long emergency is going to define who will be the winners andlosers of the planetary crisis.

And therace is already on.

Read the original here:

Artificial Intelligence, climate change and the US military - The Red (Team) Analysis Society

Artificial intelligence is among the most fascinating ideas of our time. It has captured the imagination of visionaries, science fiction writers, engineers and wall street analysts alike. In fact, artificial intelligence is in many ways a catalyst for the data revolution something that has disrupted every aspect of modern life. As with all new technologies, some are faster to embrace them, and others are much slower. Is automotive manufacturing one of the faster ones or would it be among the last?

Artificial intelligence (AI) encompasses various technologies including machine learning (ML), deep learning (neural network), computer vision and image processing, natural language processing (NLP), speech recognition, context-aware processing, and predictive APIs. But how much does this impact manufacturing and supply chain operations? Three smarts are worthy of consideration, namely smart machines, smart quality assurance and smart logistics.

The first, smart machines is relevant because improved asset utilisation is one of the greatest opportunities for AI to translate to direct savings. As overall equipment effectiveness (OEE) has been the de-facto standard to compare machine performance, automotive companies are embracing AI and machine learning (ML) algorithms to squeeze every ounce of performance from machines. Typical use cases include bottleneck detection and predictive/prescriptive maintenance. Dynamic bottleneck detection is necessary to efficiently utilise the finite manufacturing resources and to mitigate the short and long-term production constraints. In our case, we developed a neural network-based AI prediction to determine the bottleneck for the future.

A comprehensive AI strategy is vital to the success and competitiveness of automotive manufacturers, regardless of how far-fetched the use cases may seem to executives today

In terms of predictive/prescriptive maintenance, modern manufacturing machine infrastructure is designed with 3Vs for big data: volume, variability and velocity. Harnessing the potential of big data by incorporating machine learning algorithms into the data cloud, provides constant feedback to technicians and managers to ensure zero downtimes. Together with edge computing, machines are provided constant feedback based on output parameters. This leads to smarter machines that autocorrect itself based on individual cycles.

Smart quality assurance is relevant because quality controls such as quality gate are typically performed by workers. The process is often highly subjective and depends on the skill and training level of the operator. Smart assistants based on computer vision and image processing are assisting and, in some cases, taking over the inspection process. Moreover, the AI system constantly improves itself based on feedback.

The third smart is smart logistics. AI adoption in supply chains is taking off as companies realize the potential it could bring to solve their global logistic complexities, and it has a particularly significant role to play in the automotive industry.

Predictive analytics can be used to help with demand forecasting, and AI is helping network planners gain more insights on the demand patterns, resulting in improved forecasting accuracy. The efficiency gained in an accurate forecasting model has a bullwhip effect along the supply chain.

Smart warehouses are inventory systems where the inventory process is partially or entirely automated. This includes interconnected technologies to increase productivity. Smart warehouses use IIOT (Industrial Internet of Things) and AI to connect each process, data is collected at each of the nodes and the smart warehouse continuously learns and optimizes the process.

Most automakers have not taken meaningful steps towards integrating artificial intelligence in their manufacturing operations. Even the projects that do exist are mostly in partnership with universities and companies that offer products that are not customised for automotive applications.

The automotive sector, among other industries, will significantly benefit from robotic process automation (RPA) by transforming various consumer and business applications. In addition to business support functions, RPA can contribute to a number of areas in automotive manufacturing

The first movers have taken a number of initiatives (in series production, not pilot initiatives), including investments in collecting data centrally from their manufacturing operations and supply chains; projects to centrally connect a wide array of sensors to predict maintenance, uptime and other critical information using technologies such as NB-IoT; asset tracking initiatives across the supply chain; advanced predictive technologies for supply chain risks based on supplier reported KPIs and other sourced data; and investments in start-ups for predicting equipment issues.

Automotive manufacturers are often risk averse when it comes to new, unproven technologies, and it is unlikely that AI will find first application in automotive manufacturing due to a number of factors, including return on investment, which is not clear and potentially involves a protracted period; lack of expertise in AI and limited resources to dedicate to this initiative; organisational and process challenges; and availability of non-AI based approaches with satisfactory results.

Automaker manufacturing executives are interested in technology opportunities that have strong, demonstrable pay-off potential, and this is especially true in the case of suppliers. A familiar concept for the industry that has reaped rich rewards over the years is automation and robotics. Ever since the first industrial robot, the Unimate, was installed in a GM factory in 1959, automation has been one of the driving forces for the exponential growth in production and efficiency of the automotive industry. Now with hundreds of robots busy assembling parts on the manufacturing lines, a new type of robot is making waves behind the scenes to prepare for the next automotive industry revolution.

The so called softbots, or digital workforces are programmed software that can help automate many processes that are rules-driven, repetitive and involve overlapping systems. With success in HR, IT and finance, the softbots can work 24/7 on otherwise boring, repetitive manual work that normally would take days for the human workforce to complete. This could result in a significant cost reduction along with a tremendous increase in efficiency. The automotive sector, among other industries, will significantly benefit from robotic process automation (RPA) by transforming various consumer and business applications.

AI adoption in supply chains is taking off as companies realise the potential it could bring to solve their global logistic complexities, and it has a particularly significant role to play in the automotive industry

In addition to business support functions such as HR, IT, and finance, RPA can contribute to a number of areas in automotive manufacturing, including inventory management, production monitoring and balancing, paper document digitization, supplier orders and payment processing, data storage and management, and data analytics and forecasting.

RPA could take over some or most of these processes to reduce resource costs. More importantly, it can integrate with other existing technologies such as object character recognition (OCR), text mining, and nature language processing (NLP) to make more data available from the shop floor for advanced and predictive analytics. The applications can be then developed to detect or predict quality issues much faster and recommend corrective actions based on historical data and expert knowledge.

Beyond manufacturing, RPA is also making an impact in enhancing regulatory compliances such as GDPR or CCPA by helping car companies building systems to auto-process data requests by millions of users.

RPA is the next logical step and a starting point for most automotive companies. Even though RPA is rule-based and does not involve intelligence, it would help to initiate the change in mindset that is required for future AI adoption in automotive environments. In addition, RPA offers relatively quicker ROI by providing benefits in terms of cost reduction and error reduction soon after implementation.

Data-intensive manufacturing leading to data lakes, powerful computing and the availability of efficient algorithms has made it easier to integrate AI into automakers technology roadmaps. Applying AI to current manufacturing operations on a smaller scale does not require massive capital investment. Trainable data is readily available which can facilitate intensive testing and deep learning. Cloud and elastic computing have provided the opportunity to scale computing power as required. It might be beneficial to partner up with AI and ML experts from academic institutions as well as from within automaker product development teams to sustain the digital transformation journey.

Having a comprehensive AI strategy is vital to the success and competitiveness of automotive manufacturers, regardless of how far-fetched the use cases may seem to executives today.

About the authors: Anirudh Ramakrishna is Senior Consultant Industry 4.0 at umlaut; Stephen Xu and Timothy Thoppil are Managing Principals at umlaut

This article is taken from Automotive Worlds December 2019 Special report: how will artificial intelligence help run the automotive industry?,which is available now to download.

Read more:

Artificial intelligence gets to work in the automotive industry - Automotive World

December 1, 2019 Fujifilm Medical Systems U.S.A. is showcasing REiLI, the company's global medical imaging and informatics artificial intelligence (AI) technology initiative at the 2019 Radiological Society of North America's (RSNA) annual meeting.

"At RSNA 2019, we look forward to sharing the AI insights and advances we've made by working closely with clinical and research partners for several years," said Takuya Shimomura, chief technology officer and executive director, Fujifilm. "Ultimately, the long-term goal of our AI initiative is to help providers make better decisions that improve patient lives."

Under the REiLI brand, Fujifilm is developing AI technologies that strongly support diagnostic imaging workflow, leveraging the combination of its deep learning innovations and distinct image processing heritage. Applications currently in development include, but are not limited to: Region Recognition, an AI technology that helps to accurately recognize and consistently extract organ regions, regardless of deviations in shape, presence or absence of disease, and imaging conditions; Computer Aided Detection, an AI technology to reduce the time of image interpretation and support radiologists' clinical decision making; Workflow Support, using AI technology to realize optimal study prioritization, alert communications of AI findings, and report population automation.

"Our latest Synapse 7x brings diagnostic radiology, mammography and cardiology together on the server-side, enabling immediate interaction with these modality imaging data sets through a single AI-enabled platform," said Bill Lacy, vice president, medical informatics, Fujifilm. "We're excited to debut this solution for our U.S. customers at RSNA 2019, showing our commitment to progressing AI technology to empower physicians to make more efficient and impactful care decisions."

RSNA attendees are encouraged to learn more about REiLI at Booth #4111 and participate in the following Fujifilm-hosted activities.

At booth #4111, attendees can visit Fujifilm's AI Lab. The lab will feature dedicated workstations demonstrating REiLI use cases within Synapse PACS. Attendees can witness first-hand the speed and depth of the integrated workflows achieved by unifying Fujifilm's REiLI technology with the company's server-side PACS system. Featured in the AI lab will be Fujifilm developed algorithms, to include CT lung nodule, intracerebral hemorrhage, cerebral infarction MR and CT, spine label and bone temporal subtraction to name a few. In addition to the Fujifilm AI development, the AI lab will showcase its strengths by supporting a multitude of integration points in support of partner vendor and provider developed algorithms. This will include Riverain's lung nodule, MaxQ's stroke, Lunit's Chest and 2-D Mammography, LPixel's MR Aneurysm, Koios' US breast, Aidoc's pulmonary embolism and Gleamer's bone fracture.

For more inform rsna.fujimed.com

Continue reading here:

Fujifilm Showcases Artificial Intelligence Initiative And Advances at RSNA 2019 - Imaging Technology News

null

Yes, artificial intelligence (AI) is proving itself to be a worthwhile tool in the business arena at least in focused, preliminary projects. Intelligent chatbots are a classic example. Now its a question of how quickly it can be expanded to deliver on a wider basis across the business to automate decisions around inventory or investments, for example.

Theres progress on this front, as shown in McKinseys latest survey of 2,360 executives, which shows a nearly 25 percent year-over-year increase in the use of AI in various business processes and there has been a sizable jump in companies spreading AI across multiple processes.

A majority of executives in companies that have adopted AI report that it has increased revenues in areas where it is used, and 44 percent say it has reduced costs, the surveys authors, Arif Cam, Michael Chui, and Bryce Hall, all with McKinsey, state.

The results also show that a small share of companies the authors call them AI high performers are attaining outsize business results from AI. Close to two in three companies, 63 percent, report revenue increases from AI adoption in the business units. Respondents from high performers are nearly three times likelier than their lagging counterparts to report revenue gains of more than 10 percent, the survey shows.

The leading AI use cases include marketing and sales, product and service development, and supply-chain management. In marketing and sales, respondents most often report revenue increases from AI use in pricing, prediction of likelihood to buy, and customer-service analytics, the surveys authors report. In product and service development, revenue-producing use cases include the creation of new AI-based products and new AI-based enhancements. And in supply-chain management, respondents often cite sales and demand forecasting and spend analytics as use cases that generate revenue.

What are these high performers doing differently? Strategy is a key area. For example, 72 percent of respondents from AI high performers say their companies AI strategy aligns with their corporate strategy, compared with 29 percent of respondents from other companies. Similarly, 65 percent from the high performers report having a clear data strategy that supports and enables AI, compared with 20 percent from other companies. Also, the application of standardized tools to be used across the enterprise is more likely to be seen at high performers.

Adoption of Strategic AI Approaches:

Retraining workers is also a key differentiator, the survey shows. One-third of high performers, 33%, indicate the majority of their workforce has received AI-related training over the past year, compared to five percent of lagging organizations. Over the next three years, 42% of high performers intend to extend such training to most of their workers, versus only 17% of their lagging counterparts.

For AI to take hold, the McKinsey authors urge ramping up workforce retraining. Even the AI high performers have work to do in several key areas, the surveys authors point out. Only 36 percent of respondents from these companies say their frontline employees use AI insights in real time for daily decision making. A minority, 42 percent, report they systematically track a comprehensive set of well-defined key performance indicators for AI. Likewise, only 35 percent of respondents from AI high performers report having an active continuous learning program on AI for employees.

Read the original here:

It Pays To Break Artificial Intelligence Out Of The Lab, Study Confirms - Forbes