14.5 C
New York
Tuesday, March 18, 2025
Home Blog Page 3549

an OpenAI Collaboration, Generative AI, and Zero Belief


As a part of an ongoing effort to maintain you knowledgeable about our newest work, this weblog submit summarizes some latest publications from the SEI within the areas of massive language fashions for cybersecurity, software program engineering and acquisition with generative AI, zero belief, massive language fashions in nationwide safety, capability-based planning, provide chain threat administration, generative AI in software program engineering and acquisition, and quantum computing.

These publications spotlight the newest work of SEI technologists in these areas. This submit features a itemizing of every publication, creator(s), and hyperlinks the place they are often accessed on the SEI web site.

Issues for Evaluating Massive Language Fashions for Cybersecurity Duties
by Jeff Gennari, Shing-hon Lau, Samuel J. Perl, Joel Parish (OpenAI), and Girish Sastry (OpenAI)

Generative synthetic intelligence (AI) and huge language fashions (LLMs) have taken the world by storm. The power of LLMs to carry out duties seemingly on par with people has led to speedy adoption in quite a lot of totally different domains, together with cybersecurity. Nevertheless, warning is required when utilizing LLMs in a cybersecurity context because of the impactful penalties and detailed particularities. Present approaches to LLM analysis are likely to concentrate on factual information versus utilized, sensible duties. However cybersecurity duties typically require extra than simply factual recall to finish. Human efficiency on cybersecurity duties is commonly assessed partially on their capability to use ideas to real looking conditions and adapt to altering circumstances. This paper contends the identical method is important to precisely consider the capabilities and dangers of utilizing LLMs for cybersecurity duties. To allow the creation of higher evaluations, we establish key standards to contemplate when designing LLM cybersecurity assessments. These standards are additional refined right into a set of suggestions for how one can assess LLM efficiency on cybersecurity duties. The suggestions embody correctly scoping duties, designing duties based mostly on real-world cybersecurity phenomena, minimizing spurious outcomes, and making certain outcomes usually are not misinterpreted.
Learn the white paper.

The Way forward for Software program Engineering and Acquisition with Generative AI
by Douglas Schmidt (Vanderbilt College), Anita Carleton, James Ivers, Ipek Ozkaya, John E. Robert, and Shen Zhang

We stand at a pivotal second in software program engineering, with synthetic intelligence (AI) enjoying a vital function in driving approaches poised to reinforce software program acquisition, evaluation, verification, and automation. Whereas generative AI instruments initially sparked pleasure for his or her potential to cut back errors, scale modifications effortlessly, and drive innovation, issues have emerged. These issues embody safety dangers, unexpected failures, and problems with belief. Empirical analysis on generative AI improvement assistants reveals that productiveness and high quality good points rely not solely on the sophistication of instruments but in addition on activity circulate redesign and knowledgeable judgment.

On this webcast, SEI researchers discover the way forward for software program engineering and acquisition utilizing generative AI applied sciences. They look at present purposes, envision future prospects, establish analysis gaps, and talk about the essential ability units that software program engineers and stakeholders must successfully and responsibly harness generative AI’s potential. Fostering a deeper understanding of AI’s function in software program engineering and acquisition accentuates its potential and mitigates its dangers.

The webcast covers

  • how one can establish appropriate use circumstances when beginning out with generative AI expertise
  • the sensible purposes of generative AI in software program engineering and acquisition
  • how builders and resolution makers can harness generative AI expertise

View the webcast.

Zero Belief Business Days 2024 State of affairs: Secluded Semiconductors, Inc.
by Rhonda Brown

Every accepted presenter on the SEI Zero Belief Business Days 2024 occasion develops and proposes an answer for this situation: An organization is working a chip manufacturing facility on an island the place there could also be lack of connectivity and cloud companies for brief or prolonged intervals of time. There are lots of concerns when addressing the challenges of a zero belief implementation, together with various views and philosophies. This occasion provides a deep examination of how resolution suppliers and different organizations interpret and handle the challenges of implementing zero belief. Utilizing a situation locations boundaries on the zero belief area to yield richer discussions.

This yr’s occasion focuses on the Industrial Web of Issues (IIoT), legacy techniques, sensible cities, and cloud-hosted companies in a producing atmosphere.
Learn the white paper.

Utilizing Massive Language Fashions within the Nationwide Safety Realm
By Shannon Gallagher

On the request of the White Home, the Workplace of the Director of Nationwide Intelligence (ODNI) started exploring use circumstances for giant language fashions (LLMs) inside the Intelligence Group (IC). As a part of this effort, ODNI sponsored the Mayflower Mission at Carnegie Mellon College’s Software program Engineering Institute from Might 2023 by way of September 2023. The Mayflower Mission tried to reply the next questions:

  • How may the IC arrange a baseline, stand-alone LLM?
  • How may the IC customise LLMs for particular intelligence use circumstances?
  • How may the IC consider the trustworthiness of LLMs throughout use circumstances?

On this SEI Podcast, Shannon Gallagher, AI engineering crew lead, and Rachel Dzombak, former particular advisor to the director of the SEI’s AI Division, talk about the findings and proposals from the Mayflower Mission and supply further background details about LLMs and the way they are often engineered for nationwide safety use circumstances.
Hear/View the SEI Podcast.

Navigating Functionality-Primarily based Planning: The Advantages, Challenges, and Implementation Necessities
By Anandi Hira and William Nichols

Functionality-based planning (CBP) defines a framework that has an all-encompassing view of current skills and future wants for strategically deciding what is required and how one can successfully obtain it. Each enterprise and authorities acquisition domains use CBP for monetary success or to design a well-balanced protection system. The definitions understandably range throughout these domains. This paper endeavors to consolidate these definitions to offer a complete view of CBP, its potential, and sensible implementation of its rules.
Learn the white paper.

Ask Us Something: Provide Chain Danger Administration
By Brett Tucker and Matthew J. Butkovic

In response to the Verizon Knowledge Breach Report, Log4j-related exploits have occurred much less continuously over the previous yr. Nevertheless, this Frequent Vulnerabilities and Exposures (CVE) flaw was initially documented in 2021. The risk nonetheless exists regardless of elevated consciousness. Over the previous few years, the Software program Engineering Institute has developed steering and practices to assist organizations cut back threats to U.S. provide chains. On this webcast, Brett Tucker and Matthew Butkovic, reply enterprise threat administration questions to assist organizations obtain operational resilience within the cyber provide chain. The webcast covers

  • enterprise threat governance and how one can assess group’s threat urge for food and coverage because it pertains to and integrates cyber dangers into a worldwide threat portfolio
  • regulatory directives on third-party threat
  • the agenda and subjects to be lined within the upcoming CERT Cyber Provide Chain Danger Administration Symposium in February

View the webcast.

The Measurement Challenges in Software program Assurance and Provide Chain Danger Administration
by Nancy R. Mead, Carol Woody, and Scott Hissam

On this paper, the authors talk about the metrics wanted to foretell cybersecurity in open supply software program and the way requirements are wanted to make it simpler to use these metrics within the provide chain. The authors present examples of doubtless helpful metrics and underscore the necessity for information assortment and evaluation to validate the metrics. They assert that defining metrics, amassing and analyzing information for instance their utility, and utilizing normal strategies requires unbiased collaborative work to realize the specified outcomes.
Learn the white paper.

The Cybersecurity of Quantum Computing: 6 Areas of Analysis

By Tom Scanlon

Analysis and improvement of quantum computer systems continues to develop at a speedy tempo. The U.S. authorities alone spent greater than $800 million on quantum data science analysis in 2022. Thomas Scanlon, who leads the info science group within the SEI CERT Division, was not too long ago invited to be a participant within the Workshop on Cybersecurity of Quantum Computing, co-sponsored by the Nationwide Science Basis (NSF) and the White Home Workplace of Science and Know-how Coverage, to look at the rising subject of cybersecurity for quantum computing. On this SEI podcast, Scanlon discusses how one can create the self-discipline of cyber safety of quantum computing and descriptions six areas of future analysis in quantum cybersecurity.

Take heed to/view the podcast.

Product Administration Journey on the Objective Hackathon 2023 | Weblog | bol.com


As a member of Crew Paradise, our journey by way of the Objective Hackathon 2023 was an exhilarating expertise that supplied helpful insights into product administration and problem-solving. Over the course of 4 days, we engaged in a rigorous means of drawback exploration, ideation, and solution-building, with a give attention to making sustainable assortment extra engaging to prospects.

Our journey started with drawback exploration, the place we delved into the challenges of discovering a sustainable assortment. We recognized a scarcity of sustainability information hindering the inventive course of and acknowledged the necessity to have a look at the issue from completely different angles. This led us to give attention to prospects and the right way to make sustainable merchandise extra interesting to them.

Through the ideation section, we generated quite a few concepts and narrowed right down to the ultimate thought, exploring two angles: rewarding sustainability and recommending related however sustainable merchandise. We additionally performed a “5 Why’s” train, which led us to the issue assertion: “How would possibly we make a sustainable assortment really feel like deal?” This course of highlighted the significance of understanding buyer conduct and preferences, in addition to the necessity for daring and inventive approaches to deal with the challenges.

As we constructed the answer, we encountered a number of “aha” moments, resembling the conclusion that prospects merely need to really feel that their buy was deal, with high quality usually being a extra important criterion than value. We additionally realized the significance of validating assumptions with customers, as evidenced by the testing of our answer with precise bol.com customers and receiving helpful suggestions.

Our journey culminated within the creation of a service that was showcased on the bol.com stg setting. Nevertheless, we additionally confronted challenges, resembling the necessity to make the “large image” clear and be sure that the part showcasing sustainable merchandise was extra seen and fascinating for purchasers.

All through this journey, we gained a deeper understanding of the product administration course of, from drawback exploration and ideation to solution-building and consumer validation. We realized the importance of creativity, boldness, and consumer suggestions in growing options that resonate with prospects.

In conclusion, our participation within the Objective Hackathon 2023 was a transformative expertise that supplied helpful classes in product administration, innovation, and customer-centric design. We sit up for making use of these insights to future endeavors and persevering with our journey of making impactful and sustainable options for purchasers.

AI could possibly be a gamechanger for individuals with disabilities


AI might make these sorts of jumps in accessibility extra widespread throughout a variety of applied sciences. However you in all probability haven’t heard a lot about that risk. Whereas the New York Instances sues OpenAI over ChatGPT’s scraping of its content material and everybody ruminates over the ethics of AI instruments, there appears to be much less consideration of the good ChatGPT can do for individuals of assorted talents. For somebody with visible and motor delays, utilizing ChatGPT to do analysis is usually a lifesaver. As an alternative of making an attempt to handle a dozen browser tabs with Google searches and different pertinent data, you’ll be able to have ChatGPT collate every thing into one area. Likewise, it’s extremely believable that artists who can’t draw within the typical method might use voice prompts to have Midjourney or Adobe Firefly create what they’re pondering of. That is likely to be the one method for such an individual to indulge an inventive ardour. 

For many who, like me, are blind or have low imaginative and prescient, the flexibility to summon a experience on demand and go anyplace with out imposing on anybody else for assist is a big deal.

After all, knowledge must be vetted for accuracy and gathered with permission—there are ample causes to be cautious of AI’s potential to serve up improper or probably dangerous, ableist details about the disabled neighborhood. Nonetheless, it feels unappreciated (and underreported) that AI-based software program can really be an assistive know-how, enabling individuals to do issues they in any other case can be excluded from. AI might give a disabled particular person company and autonomy. That’s the entire level of accessibility—liberating individuals in a society not designed for his or her wants.

The flexibility to mechanically generate video captions and picture descriptions supplies further examples of how automation could make computer systems and productiveness know-how extra accessible. And extra broadly, it’s laborious to not be enthused about ever-burgeoning applied sciences like autonomous autos. Most tech journalists and different business watchers are eager about self-driving vehicles for the sheer novelty, however the actuality is the AI software program behind autos like Waymo’s fleet of Jaguar SUVs is sort of actually enabling many within the incapacity neighborhood to exert extra company over their transport. For many who, like me, are blind or have low imaginative and prescient, the flexibility to summon a experience on demand and go anyplace with out imposing on anybody else for assist is a big deal. It’s not laborious to ascertain a future during which, because the know-how matures, autonomous autos are normalized to the purpose the place blind individuals might purchase their very own vehicles. 

Google I/O 2024: Shared Factor Transitions in Jetpack Compose


Google I/O 2024 launched so many thrilling new applied sciences, particularly Gemini AI and Jetpack Compose. Although Jetpack Compose is superior and bettering quickly to meet up with the legacy XML-based layouts (which have been on the market for ages), it fell quick in some areas, resembling animations.

Shared Factor Transitions are among the many hottest APIs from the Android Animation Framework, which wasn’t accessible within the Jetpack Compose till now. That’s proper – Google launched shared ingredient transitions for Jetpack Compose at their 2024 I/O occasion!

This long-awaited characteristic helps you create stunning, fluid animations when navigating between screens in your app. Think about a person tapping a picture in an inventory, and it easily expands and animates into the detailed view. Shared ingredient transitions in Compose present a declarative approach to obtain this impact, supplying you with extra management over the animation course of than the standard View system. This empowers builders to design seamless person experiences that improve their apps’ general feel and look.

Listed here are a number of the key capabilities of Shared Factor Transitions in Jetpack Compose launched at Google I/O 2024:

  • Declarative Animation: Shared ingredient transitions are outlined declaratively utilizing modifiers like Modifier.sharedElement and Modifier.sharedBoundsMatchingContentSize. This animation course of is far less complicated in comparison with the crucial method required within the View system.
  • Finer Management: Compose offers extra granular management over the animation in comparison with conventional strategies. You possibly can outline the particular ingredient to animate, its transition bounds, and even the animation kind.
  • Seamless Integration With Navigation: Shared ingredient transitions work easily with Navigation Compose. When navigating between screens, you may cross the ingredient’s key as an argument, and Compose routinely matches components and creates the animation.

Getting Began

To benefit from the newest APIs, be sure you’re utilizing the newest Android Studio Jellyfish | 2023.3.1 and API Stage 34.

Click on the Obtain Supplies button on the prime or backside of this tutorial. Unzip the ComposeTransitionAnimation.zip folder.

Now, launch Android Studio and open ComposeTransitionAnimation-Starter to import the starter mission. The ComposeTransitionAnimation-Starter mission accommodates the required boilerplates and Composables to leap straight into the animation!

ComposeTransitionAnimation-Starter resembles an e-commerce app with a primary Record-Element format.

Construct and run the app – it’ll appear like this:

On this article, you’ll create a visible connection between components on Record and Element screens utilizing Shared Factor Transition.

First, add the newest model of Compose dependencies. Open construct.gradle in your app module and replace:

def composeVersion = "1.7.0-beta01"

Faucet Sync Now to obtain the dependencies.

Notice: Shared ingredient help is experimental and is in `beta`. The APIs could change sooner or later.

Overview of Key APIs

The most recent dependencies launched just a few high-level APIs that do the heavy lifting of sharing components between Composable layouts:

  • SharedTransitionLayout: The highest-level format required to implement shared ingredient transitions. It offers a SharedTransitionScope. A Composable must be in SharedTransitionScope to make use of the modifiers of shared components.
  • Modifier.sharedElement(): The modifier to flag one Composable to be matched with one other Composable inside the SharedTransitionScope.
  • Modifier.sharedBounds(): The modifier that tells the SharedTransitionScope to make use of this Composable’s bounds because the container bounds for the place the transition ought to happen.

You’ll quickly create a hero-animation utilizing these APIs.

Implementing Shared Transition Animation

A Shared Transition Animation, or hero-animation, consists of three main steps:

  1. Wrapping collaborating views with SharedTransitionLayout.
  2. Defining SharedTransitionScope to the supply and vacation spot views.
  3. Transition with Shared Factor.

Including SharedTransitionLayout

Open the MainActivity class. It accommodates ListScreen and DetailScreen, which can share components throughout a transition animation. As talked about earlier, you should wrap them with SharedTransitionLayout to make them eligible for a Shared Transition Animation.

Replace the AnimatedContent block as follows:

SharedTransitionLayout {
  AnimatedContent(
    targetState = showDetails, 
    label = "shared_transition"
  ) { shouldShowDetails ->
    if (!shouldShowDetails) {
      ListScreen(
        // Present code
        ... ... ...
      )
    } else {
      DetailScreen(
        // Present code
        ... ... ...
       )
     }
  }
}

At this level, you may even see this warning from Android Studio for utilizing an experimental api:

To resolve this, add these imports on prime of the MainActivity:

import androidx.compose.animation.ExperimentalSharedTransitionApi
import androidx.compose.animation.SharedTransitionLayout

Then add this annotation over the onCreate(savedInstanceState: Bundle?) methodology:

@OptIn(ExperimentalSharedTransitionApi::class)

Construct and run.

Defining SharedTransitionScope

Up subsequent, you should outline SharedTransitionScope to the views collaborating within the transition animation. The Composable must be inside SharedTransitionScope to make use of Modifier.sharedElement() for the animation. Therefore, you’ll have to cross down SharedTransitionScope from SharedTransitionLayout in MainActivity to the supply and vacation spot Composable executing the animation.

On this case, you’ll transition from the smaller Picture Composable within the ListScreen (supply) to the bigger Composable in DetailScreen (vacation spot).

Begin with ListScreen.kt inside ui package deal. Replace the ListScreen perform with these parameters:

@Composable
enjoyable ListScreen(
  paddingValues: PaddingValues,
  gadgets: Record,
  onItemClicked: (Merchandise) -> Unit = {},
  sharedTransitionScope: SharedTransitionScope,
  animatedVisibilityScope: AnimatedVisibilityScope,
)

Then cross the sharedTransitionScope and animatedVisibilityScope references for every ListItem:

gadgets.forEach { merchandise ->
  ListItem(
    merchandise = merchandise,
    onItemClicked = onItemClicked,
    sharedTransitionScope = sharedTransitionScope,
    animatedVisibilityScope = animatedVisibilityScope,
  )
}

Additionally, replace th eListItem Composable methodology signature accordingly:

@Composable
enjoyable ListItem(
  merchandise: Merchandise,
  onItemClicked: (Merchandise) -> Unit = {},
  sharedTransitionScope: SharedTransitionScope,
  animatedVisibilityScope: AnimatedVisibilityScope,
)

You’ll see the warning for utilizing an experimental api once more from the compiler, together with the errors for the lacking imports.

Fret not! Add these imports on prime:

import androidx.compose.animation.AnimatedVisibilityScope
import androidx.compose.animation.ExperimentalSharedTransitionApi
import androidx.compose.animation.SharedTransitionScope

And the annotation for the ListScreen.kt file, above of the package deal title like this:

@file:OptIn(ExperimentalSharedTransitionApi::class)

package deal com.kodeco.android.composetransition.ui

That ensures you may have all the required imports and can mute warnings for utilizing experimental APIs for the scope of the ListScreen.kt file.

Notice: Add the imports and annotation on DetailScreen.kt, too. You’ll want them shortly!

Your vacation spot Composable is the DetailScreen methodology. Now add animation scopes as methodology parameters as follows:

@Composable
enjoyable DetailScreen(
  merchandise: Merchandise, onBack: () -> Unit,
  sharedTransitionScope: SharedTransitionScope,
  animatedVisibilityScope: AnimatedVisibilityScope,
)

You’re able to wire up ListScreen and DetailScreen to carry out the transition animation.

Open MainActivity and replace SharedTransitionLayout block to cross animatedVisibilityScope and sharedTransitionScope to its descendants:

SharedTransitionLayout {
  AnimatedContent(
    targetState = showDetails, 
    label = "shared_transition"
  ) { shouldShowDetails ->
    if (!shouldShowDetails) {
      ListScreen(
        paddingValues = paddingValues,
        gadgets = gadgets.worth,
        onItemClicked = { merchandise ->
          detailItem = merchandise
          showDetails = !showDetails
        },
        animatedVisibilityScope = this@AnimatedContent,
        sharedTransitionScope = this@SharedTransitionLayout,
      )
    } else {
      DetailScreen(
        merchandise = detailItem,
        onBack = { showDetails = !showDetails },
        animatedVisibilityScope = this@AnimatedContent,
        sharedTransitionScope = this@SharedTransitionLayout,
      )
    }
  }
}

Construct and run once more to make sure you resolved all compilation errors, however don’t count on the animation to occur but!

New Home windows 10 22H2 beta fixes reminiscence leaks and crashes

0


New Home windows 10 22H2 beta fixes reminiscence leaks and crashes

Microsoft has launched a brand new Home windows 10 22H2 beta (KB5041582) with reminiscence leak and crash fixes for Insiders within the Beta and Launch Preview channels.

The corporate says the KB5041582 replace fixes a difficulty inflicting some apps to cease due to a reminiscence leak in a Bluetooth system and an Enter Methodology Editor (IME) reminiscence leak occurring when closing combo packing containers with enter focus.

It additionally addressed a Bind Filter Driver bug that brought on impacted programs to freeze when accessing symbolic hyperlinks and a restart difficulty triggered by a Microsoft System Middle Configuration Supervisor (SCCM) job not having the ability to re-enable Unified Write Filter (UWF) due to a impasse.

“The Beta Channel and Launch Preview Channel will obtain the identical Home windows 10 builds; nevertheless, the Beta Channel is the place we’ll attempt new options out with Insiders that we’re engaged on that aren’t but prepared for all Home windows 10 clients,” the Home windows Insider Program Workforce stated on Thursday.

KB5041582 is now obtainable for Insider programs operating Home windows 10, model 22H2. As soon as put in, it should replace the model to Construct 19045.4842.

Home windows 10 Beta channel reopened after three years

Microsoft reopened the Home windows 10 beta channel in early June and introduced the primary Home windows 10 Beta construct since 2021, three years after the final such construct was launched for Insiders within the Beta and Launch Preview channels.

“To carry new options and extra enhancements to Home windows 10 as wanted, we want a spot to do lively characteristic growth with Home windows Insiders,” the Home windows Insider staff stated on June 4.

“So at present, we’re opening the Beta Channel for Home windows Insiders who’re presently on Home windows 10. It will permit us to check out new options for Home windows 10, model 22H2, with Insiders earlier than releasing them to all Home windows 10 clients.”

Insiders who wish to check new Home windows 10 options and fixes as quickly as they’re obtainable ought to be a part of or change to the Beta Channel by going to Settings > Replace & Safety > Home windows Insider Program and selecting the Beta channel after clicking on the present Insider channel choice to get to the settings web page and the listing of obtainable Insider channels.

Nevertheless, switching to the Canary or Dev channels will improve their gadgets to the newest Home windows 11 construct obtainable for these channels. It is also necessary to notice that Microsoft won’t instantly improve the programs and provides customers a short interval to roll again to Home windows 10, which would require a clear set up.