generates Archives - Business Intelligence Info

Tag Archives: generates

LinkedIn’s AI generates candidate screening questions from job postings

May 1, 2020 Big Data

LinkedIn is using AI and machine learning to generate screening questions for active job postings. In a paper published this week on the preprint server Arxiv.org, coauthors describe Job2Questions, a model that helps recruiters quickly find applicants by reducing the need for manual screening. This isn’t just theoretical research — Job2Questions was briefly tested across millions of jobs by hiring managers and candidates on LinkedIn’s platform.

The timing of Job2Questions’ deployment is fortuitous. Screening is a necessary evil — a LinkedIn study found that roughly 70% of manual phone screenings uncover missing basic applicant qualifications. But as the pandemic increasingly impacts traditional hiring processes, companies are adopting alternatives, with some showing a willingness to pilot AI and machine learning tools. Job2Questions is designed to reduce the time recruiters spend asking questions they should already have answers to or exposes gaps candidates themselves can fill.

As the researchers explain, Job2Questions generates a number of screening question candidates, given the content of a job posting. It first divides postings into sentences and converts these sentences into pairs of question templates (e.g., “How many years of work experience do you have using…” and “Have you completed the following level of education:”) and variables (“Java” and “Bachelor’s Degree”). Then, it classifies the sentences into one of several templates designed by hiring experts and taps an entity linking system to detect the parameters corresponding to the chosen templates, namely by tagging specific types of entities from the sentences (like “education degrees,” “spoken languages,” “tool-typed skills,” and “credentials”). A pretrained, fine-tuned deep averaging network within Job2Questions parses posting text for semantic meaning. And lastly, a ranking model identifies the best questions of the bunch.

To collect data to train the machine learning models underpinning Job2Questions, the LinkedIn researchers had annotators label sentence-question pairs, which enabled the prediction of the templates from sentences. Then, the team collected 110,409 labeled triples — data samples containing a single job posting, a template, and parameters — submitted by job posters on LinkedIn, which served to train Job2Questions’ question-ranking model to anticipate whether a job poster would add a screening question to a posting. Screening questions added and rejected by recruiters and posters served as ground-truth labels.

VB Transform 2020 Online – July 15-17, 2020: Join leading AI executives at VentureBeat’s AI event of the year. Register today and save 30% off digital access passes.

In the course of a two-week experiment involving 50% of LinkedIn’s traffic, the researchers claim that only 18.67% of applicants who didn’t answer screening questions correctly were rated as a “good fit” by recruiters, while those who answered at least one question correctly had a 23% higher ranking. They also claim that ranking candidates by their screening question answers improved the applicant good fit rate by 7.45% and reduce the bad fit rate by 1.67%; that applicants were 46% more likely to get a good fit rating for job recommendations informed by their answers to questions; and that jobs with screening questions yielded 1.9 times more recruiter-applicant interaction in general and 2.4 times more interactions with screening-qualified applicants.

“We found that screening questions often contains information that members do not put in their profile. Among members who answered screening questions, 33% of the members do not provide their education information in their profile. More specifically, people who hold secondary education degree are less likely to list that in their profile. As for languages, 70% of the members do not list the languages they spoke (mostly native speakers) in their profile. Lastly, 37% of the members do not include experience with specific tools,” wrote the paper’s coauthors. “In short, we suspect that when people [are] composing their professional profile, they tend to overlook basic qualifications which recruiters value a lot during screening. Therefore, screening questions are much better, direct signals for applicant screening compared to member profile.”

Let’s block ads! (Why?)

Big Data – VentureBeat

IBM’s AI generates new footage from video stills

March 25, 2020 Big Data

A paper coauthored by researchers at IBM describes an AI system — Navsynth — that generates videos seen during training as well as unseen videos. While this in and of itself isn’t novel — it’s an acute area of interest for Alphabet’s DeepMind and others — the researchers say the approach produces superior quality videos compared with existing methods. If the claim holds water, their system could be used to synthesize videos on which other AI systems train, supplementing real-world data sets that are incomplete or marred by corrupted samples.

As the researchers explain, the bulk of work in the video synthesis domain leverages GANs, or two-part neural networks consisting of generators that produce samples and discriminators that attempt to distinguish between the generated samples and real-world samples. They’re highly capable but suffer from a phenomenon called mode collapse, where the generator generates a limited diversity of samples (or even the same sample) regardless of the input.

By contrast, IBM’s system consists of a variable representing video content features, a frame-specific transient variable (more on that later), a generator, and a recurrent machine learning model. It breaks videos down into a static constituent that captures the constant portion of the video common for all frames and a transient constituent that represents the temporal dynamics (i.e., periodic regularity driven by time-based events) between all the frames in the video. Effectively, the system jointly learns the static and transient constituents, which it uses to generate videos at inference time.

IBM’s AI generates new footage from video stills

Above: Videos trained by IBM’s Navsynth system.

To capture equally from the static portion of the video, the researchers’ system randomly chooses a frame and compares its corresponding generated frame during training. This ensures that the generated frame remains close to the ground truth frame.

In experiments, the research team trained, validated, and tested the system on three publicly available data sets: Chair-CAD, which consists of 1,393 3D models of chairs (out of which 820 were chosen with the first 16 frames); Weizmann Human Action, which provides 10 different actions performed by 9 people, amounting to 90 videos; and the Golf scene data set, which contains 20,268 golf videos (out of which 500 videos were chosen).

The researchers say that, compared with the videos generated by several baseline models, their system produced “visually more appealing” videos that “maintained consistency” with sharper frames. Moreover, it reportedly demonstrated a knack for frame interpolation, or a form of video processing in which the intermediate frames are generated between the existing on in an attempt to make animation more fluid.

Let’s block ads! (Why?)

Big Data – VentureBeat

Microsoft’s AI generates 3D objects from 2D images

March 6, 2020 Big Data

The AI research labs at Facebook, Nvidia, and startups like Threedy.ai have at various points tried their hand at the challenge of 2D-object-to-3D-shape conversion. But in a new preprint paper, a team hailing from Microsoft Research detail a framework that they claim is the first “scalable” training technique for 3D models from 2D data. They say it can consistently learn to generate better shapes than existing models when trained with exclusively 2D images, which could be a boon for video game developers, ecommerce businesses, and animation studios that lack the means or expertise to create 3D shapes from scratch.

In contrast to previous work, the researchers sought to take advantage of fully featured industrial renderers — i.e., software that produces images from display data. To that end, they train a generative model for 3D shapes such that rendering the shapes generates images matching the distribution of a 2D data set. The generator model takes in a random input vector (values representing the data set’s features) and generates a continuous voxel representation (values on a grid in 3D space) of the 3D object. Then, it feeds the voxels to a non-differentiable rendering process, which thresholds them to discrete values before they’re rendered using an off-the-shelf renderer (the Pyrender, which is built on top of OpenGL).

A novel proxy neural renderer directly renders the continuous voxel grid generated by the 3D generative model. As the researchers explain, it’s trained to match the rendering output of the off-the-shelf renderer given a 3D mesh input.

Microsoft’s AI generates 3D objects from 2D images

Above: Couches, chairs, and bathtubs generated by Microsoft’s model.

Image Credit: Microsoft

In experiments, the team employed a 3D convolutional GAN architecture for the generator. (GANs are two-part AI models comprising generators that produce synthetic examples from random noise sampled using a distribution, which along with real examples from a training data set are fed to the discriminator, which attempts to distinguish between the two.) Drawing on a range of synthetic data sets generated from 3D models and a real-life data set, they synthesized images from different object categories, which they rendered from different viewpoints throughout the training process.

Above: Mushrooms generated by the model.

Image Credit: Microsoft

The researchers say that their approach takes advantage of the lighting and shading cues the images provide, enabling it to extract more meaningful information per training sample and produce better results in those settings. Moreover, it’s able to produce realistic samples when trained on data sets of natural images. “Our approach … successfully detects the interior structure of concave objects using the differences in light exposures between surfaces,” wrote the paper’s coauthors, “enabling it to accurately capture concavities and hollow spaces.”

They leave to future work incorporating color, material, and lighting prediction into their system to extend it to work with more “general” real-world data sets.

Let’s block ads! (Why?)

Big Data – VentureBeat

Tips for Creating a Landing Page That Generates Conversions

February 27, 2020 CRM News and Info

by Helen Veyna

Last updated February 19, 2020

If demand and lead generation is a top priority for your organization, you know that capturing critical contact information is key to making sure your leads don’t slip away. Knowing who your contacts are, what industry they work in, and what is driving them to explore your solution empowers you to continue to engage with them through personalized marketing efforts. After all, the point of having a robust demand generation program is to attract leads that you can nurture and convert into customers down the road.

shutterstock 218538049 Tips for Creating a Landing Page That Generates Conversions

Making sure that your messaging and content are right on target will help you build credibility with these leads, keep them moving through the sales funnel, and convince them to choose you over your competitors.

But you’re probably wondering how to go about collecting key information, especially when it’s challenging enough to even get people on your website.

In some cases, landing pages allow you to kill two birds with one stone. The right landing page will help drive people to your site by offering original content and information and offer these individuals valuable content gated behind forms to entice them to hand over their information. However, if your goal is to generate conversions, you can use dedicated landing pages without any extra navigation to focus on the goal at hand.

The information you collect using landing pages allows you to build a thorough customer profile to help you determine the best marketing strategy to keep each customer interested. As you uncover more information about who your customers are, you can segment them into the appropriate campaigns and send them the right content and information to help them inch closer toward making a decision.

It’s important to keep in mind, however, that not all landing pages are created equal. Having a well-designed landing page can help you stand out from the competition and generate conversions. Today, we’re outlining a few best practices to help you make sure you’re creating landing pages that are visually appealing, easy to read, and motivate your customers to complete your call to action.

1. Make Sure Your Headline Is Clear and Catchy

Your headline is the first thing that people see when visiting your landing page. Many times, that’s all they need to decide if they want to continue to engage. It can also be extremely disappointing for visitors to visit a webpage that doesn’t offer what it said it would. So, it’s extremely important for your headlines to not only be catchy, but also clearly state what you have to offer.

If the headline is the title of a content asset (such as an eBook), you should use titles that directly address the kind of information your visitors can expect to find. If you need a little more room to make your offer easy to understand, you can always use a subheadline to further explain the purpose of your landing page. In addition to providing your visitors a clear idea of the offer they can expect from your landing page, your headline also provides a great opportunity to use popular keywords that resonate with your audience and will drive more traffic to your landing pages.

2. Keep Your Copy Brief and to the Point

A landing page should serve as an introduction of what’s in store for your visitors if they decide to exchange their information. Having too much copy on a page can overwhelm the user and make them bounce from your landing page before filling out a form.

To avoid this scenario, keep your copy brief and to the point. A short paragraph and a few bullet points is all you need to provide your audience with a sneak peek of what they’ll get when they fill out a form and download your asset or sign up for an event.

Because you’re keeping your copy short, you need to ensure that your message is clear and resonates with the pain points and interests of your consumers. Using high-ranking keywords and language that pertains to your visitors industry, preferences, and stage in the sales funnel can help ensure that your prospects remain interested in what you have to say.

rq5Kvcsd7okTUEeCmcVPNn Tips for Creating a Landing Page That Generates Conversions

3. Always Include Important Information

If you’re using your landing page to promote a piece of content such as an eBook or video, you can get away with including a few bullet points describing what your visitors can expect. If you’re hosting an event or webinar, however, you should also include a date, time, location, and presenter information.

Doing this will help your leads remember to clear time on their schedule to attend, helping you reap the rewards of interacting with in real time.

4. Use Design Best Practices to Make Your Landing Page Visually Appealing

A landing page that is too busy and hard to read will hurt your conversion rate. If you want to prevent your audience from bouncing before converting, you need good and simple design as a foundation.

Keeping things simple doesn’t mean that your landing page should be plain. Using eye catching graphics and images that complement the purpose of the landing page and give more context to your copy are a great way to capture your audience’s attention.

Another good rule of thumb is to keep key information above the fold. While you can include information on your landing page that allows your visitors to learn more about your brand and what you have to offer, you don’t want to distract them from your call to action. Keeping elements such as summaries and CTA buttons at the top of the page will help your visitors know what to do to receive their offer and make it easy for them to do so.

You don’t have to be a graphic designer or know HTML to design a visually appealing landing page. A marketing automation platform with easy-to-use templates will have a variety of features to help you get started.

Building Better Landing Pages Banner CTA Tips for Creating a Landing Page That Generates Conversions

5. Make It Easy for Individuals to Sign Up

A form with too many fields will scare away potential leads. Using an adaptive form with minimal fields allows you to ask customers for only the information you need to continue to build their profile. So, for example, if you already have their email and name, you can ask them to indicate their industry and topic of interest. This helps you collect the information you need and eliminates your lead’s frustration of having to fill out the same information each and every time.

6. Include a Clear CTA That Is Easy to Find

For your landing page to generate conversions, you have to make sure that you have a clear call to action. Whether it’s downloading an eBook, registering for an event, or signing up to receive more information, your customers have to know what to do and how to do it. Placing a prominent button at the end of your form will indicate that you want your visitors to fill out a form — and what they’ll get if they do it.

7. Feature Key Customer Logos and Reviews

Many times, a prospective customer’s first impression of your brand is your landing page. So, you want to show your audience who you are (beyond an eBook or a webinar). An easy way to do that is by showcasing a few top customer reviews and notable logos at the bottom of your page. This will allow your prospects to gain a sense of the type of customers you serve and learn why they choose you over your competitors, while also helping you establish credibility and trust in what you do.

Tips for Creating a Landing Page That Generates Conversions

8. A/B Test Elements to Optimize Your Efforts

Even the best landing pages can use a refresh every now and then. That is why you should make A/B testing different elements of your page a priority. A good place to start is by testing different versions of your headline to see if it’s helping drive traffic to your page. Once that is optimized, you can venture on to test other elements — such as design, CTA placement, and so on.

Act-On Can Help You Build Engaging Landing Pages That Spark the Customer Journey

Landing pages are one of the most powerful tools at your disposal because they allow you to collect that information you need to deliver personalized marketing efforts. But creating compelling and effective landing pages shouldn’t require you to invest more time and resources. Act-On makes it easy to build engaging landing pages and adaptive forms that appeal to your audience and generate conversions.

Additionally, our platform can empower you to do so much more than build great landing pages. We give you the tools to launch innovative and personalized multi-channel campaigns that help you attract leads, convert customers, and improve retention. It’s your true one-stop shop for lifecycle marketing!

To learn how Act-On can help you transform your digital marketing strategy, please schedule a demo with one of our marketing automation experts. They’ll be thrilled to show you the many ways Act-On makes it easy to innovate your marketing strategy at scale.

Building Better Landing Pages blog CTA Tips for Creating a Landing Page That Generates Conversions

MIT CSAIL’s TextFooler generates adversarial text to strengthen natural language models

February 9, 2020 Big Data

AI and machine learning algorithms are vulnerable to adversarial samples that have alterations from the originals. That’s especially problematic as natural language models become capable of generating humanlike text, because of their attractiveness to malicious actors who would use them to produce misleading media. In pursuit of a technique that illustrates the extent to which adversarial text can affect model prediction, researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), the University of Hong Kong, and Singapore’s Agency for Science, Technology, and Research developed TextFooler, a baseline framework for synthesizing adversarial text examples. They claim in a paper that it was able to successfully attack three leading target models, including Google’s BERT.

“If those tools are vulnerable to purposeful adversarial attacking, then the consequences may be disastrous,” said Di Jin, MIT Ph.D. student and lead author on the paper, who noted that the adversarial examples produced by TextFooler could improve the robustness of AI models trained on them. “These tools need to have effective defense approaches to protect themselves, and in order to make such a safe defense system, we need to first examine the adversarial methods.”

The researchers assert that besides the ability to fool AI models, the outputs of a natural language “attacking” system like TextFooler should meet certain criteria: human prediction consistency, such that human predictions remain unchanged; semantic similarity, such that crafted examples bear the same meaning as the source; and language fluency, such that generated examples look natural and grammatical. TextFooler meets all three even when no model architecture or parameters (values that influence model performance) are available — i.e., black-box scenarios.

It achieves this by identifying the most important words for the target models and replacing them with semantically similar and grammatically correct words until the prediction is altered. TextFooler is applied to two different tasks — text classification and entailment (the relationship between text fragments in a sentence) — with the goal of changing the classification or invalidating the entailment judgment of the original models. For instance, given the input “The characters, cast in impossibly contrived situations, are totally estranged from reality,” TextFooler might output “The characters, cast in impossibly engineered circumstances, are fully estranged from reality.”

To evaluate TextFooler, the researchers applied it to text classification data sets with various properties, including news topic classification, fake news detection, and sentence- and document-level sentiment analysis, where the average text length ranged from tens of words to hundreds of words. For each data set, they trained the aforementioned state-of-the-art models on a training set before generating adversarial examples semantically similar to the test set to attack those models.

The team reports that on the adversarial examples, they managed to reduce the accuracy of almost all target models in all tasks to below 10% with fewer than 20% of the original words perturbed. Even for BERT, which attained relatively robust performance compared with the other models tested, TextFooler reduced its prediction accuracy by about 5 to 7 times on a classification task and about 9 to 22 times on an entailment task (where the goal was to judge whether a sentence could be derived from entailment, contradiction, or a neutral relationship).

“The system can be used or extended to attack any classification-based NLP models to test their robustness,” said Jin. “On the other hand, the generated adversaries can be used to improve the robustness and generalization of deep learning models via adversarial training, which is a critical direction of this work.”

Let’s block ads! (Why?)

Big Data – VentureBeat

IBM’s biology-inspired AI generates hash codes faster than classical approaches

January 22, 2020 Big Data

Ever heard of FlyHash? It’s an algorithm inspired by fruit flies’ olfactory circuits that’s been shown to generate hash codes — numeric representations of objects — with performance superior to classical algorithms. Unfortunately, because FlyHash uses random projections, it can’t learn from data. To overcome this limitation, researchers at Princeton, the University of San Diego, IBM Research, and the MIT-IBM Watson AI Lab developed BioHash, which applies “local” and “biologically plausible” synaptic plasticity rules to produce hash codes. They say that it outperforms previously published benchmarks for various hashing methods and that it could yield binary representations of things useful for similarity searches.

As the researchers explain in a preprint paper detailing their work, the phenomenon known as expansive representation is nearly ubiquitous in neurobiology. “Expansion” in this context refers to the mapping of high-dimensional input data to an even higher-dimensional secondary representation. For instance, in the abovementioned fruit fly olfactory system, approximately 50 neurons send their activities to about 2,500 cells called Kenyon cells, achieving an approximately 50 times expansion.

From a computational perspective, expansion can among other things increase the memory storage capacity of an AI model. It’s with this motivation that the team designed the hashing algorithm BioHash, which can be used in similarity search.

In similarity search, given a query, a similarity measure, and a database containing any number of items, the objective is to retrieve a ranked list of items from the database most similar to the query. When the data is high-dimensional (e.g. images or documents) and the databases are large (in the millions or billions of items), it’s a computationally challenging problem. However, approximate solutions are generally acceptable, including a hashing scheme called locality-sensitive hashing (LHS) in which each database entry is encoded with binary representations and retrieves closely related entries.

FlyHash leverages LHS, as does BioHash. But importantly, BioHash is faster and much more scalable.

The researchers trained and tested Biohash on MNIST, a data set of 70,000 grayscale images of handwritten digits with 10 classes of digits ranging from “0” to “9”, and CIFAR-10, a corpus comprising 60,000 images from 10 classes (e.g., “car,” “bird”). They say that BioHash demonstrated the best retrieval performance in terms of speed, substantially outperforming other methods, and that a refined version of BioHash — BioConvHash — performed even better thanks to the incorporation of purpose-built filters.

The team asserts that this provides evidence that the reason expansive representations are common in living things is because they perform LHS. In other words, they cluster similar stimuli together and push distinct stimuli far apart. “[Our] work provides evidence toward the proposal that LHS might be a fundamental computational principle utilized by the sparse expansive circuits … [Biohash] produces sparse high dimensional hash codes in a data-driven manner and with learning of synapses in a neurobiologically plausible way.”

As it turns out, the fields of neurobiology and machine learning go hand in hand. Google parent company Alphabet’s DeepMind earlier this month published a paper investigating whether the brain represents possible future rewards not as a single average but as a probability distribution, a mathematical function that provides the probabilities of occurrence of different outcomes. And scientists at Google and the Max Planck Institute of Neurobiology recently demonstrated a recurrent neural network — a type of machine learning algorithm that’s often used in handwriting and speech recognition — that maps the brain’s neurons.

Let’s block ads! (Why?)

Big Data – VentureBeat

IBM’s Lambada AI generates training data for text classifiers

November 15, 2019 Big Data

What’s a data scientist to do if they lack sufficient data to train a machine learning model? One potential avenue is synthetic data generation, which researchers at IBM Research advocate in a newly published preprint paper. They used a pretrained machine learning model to artificially synthesize new labeled data for text classification tasks. They claim that their method, which they refer to as language-model-based data augmentation (Lambada for short), improves classifiers’ performance on a variety of data sets and significantly improves upon state-of-the-art techniques for data augmentation.

“Depending upon the problem at hand, getting a good fit for a classifier model may require abundant labeled data. However, in many cases, and especially when developing AI systems for specific applications, labeled data is scarce and costly to obtain,” wrote the paper’s coauthors. “Depending upon the problem at hand, getting a good fit for a classifier model may require abundant labeled data. However, in many cases, and especially when developing AI systems for specific applications, labeled data is scarce and costly to obtain.”

Generating synthetic training data tends to be more challenging in the text domain than the visual domain, the researchers note, because the transformations used in simpler methods usually distort the text, making it grammatically and semantically incorrect. That’s why most text data augmentation techniques — including those detailed in the paper — involve replacing a single word with a synonym, deleting a word, or changing the word order.

Lambada leverages a generative model (OpenAI’s GPT) that’s pretrained on large bodies of text, enabling it to capture the structure of language such that it produces coherent sentences. The researchers fine-tuned their model on an existing, small data set, and used the fine-tuned model to synthesize new labeled sentences. Independently, they trained a classifier on the aforementioned data set and had it filter the synthesized corpus, retaining only data that appeared to be “qualitative enough” before re-training the classifier on both the existing and synthesized data.

To validate their approach, the researchers tested three different classifiers — BERT, a support vector machine, and a long short-term memory network — on three data sets by running experiments in which they varied the training samples per class. The corpora in question contained queries on flight-related information, open-domain and fact-based questions in several categories, and data from telco customer support systems.

They report that Lambada statically improved all three classifiers’ performance on small data sets, which they attribute in part to its controls over the number of samples per class. Said controls allowed them to invest more time in generating samples for classes that are under-represented in the original data set, they said.

“Our augmentation framework does not require additional unlabeled data … Surprisingly, for most classifiers, LAMBADA achieves better accuracy compared to a simple weak labeling approach,” wrote the coauthors. “Clearly, the generated data set contributes more to improving the accuracy of the classifier than … samples taken from the original data set.”

Let’s block ads! (Why?)

Big Data – VentureBeat

Microsoft’s AI generates high-quality talking heads from audio

October 8, 2019 Big Data

A growing body of research suggests that the facial movements of almost anyone can be synced to audio clips of speech, given a sufficiently large corpus. In June, applied scientists at Samsung detailed an end-to-end model capable of animating the eyebrows, mouth, and eyelashes, and cheeks in a person’s headshot. Only a few weeks later, Udacity revealed a system that automatically generates standup lecture videos from audio narration. And two years ago, Carnegie Mellon researchers published a paper describing an approach for transferring the facial movements from one person to another.

Building on this and other work, a Microsoft Research team this week laid out a technique they claim improves the fidelity of audio-driven talking heads animations. Previous head generation approaches required clean and relatively noise-free audio with a neutral tone, but the researchers say their method — which disentangles audio sequences into factors like phonetic content and background noise — can generalize to noisy and “emotionally rich” data samples.

“As we all know, speech is riddled with variations. Different people utter the same word in different contexts with varying duration, amplitude, tone and so on. In addition to linguistic (phonetic) content, speech carries abundant information revealing about the speaker’s emotional state, identity (gender, age, ethnicity) and personality to name a few,” explained the coauthors. “To the best of our knowledge, [ours] is the first approach of improving the performance from audio representation learning perspective.”

Underlying their proposed technique is a variational autoencoder (VAE) that learns latent representations. Input audio sequences are factorized by the VAE into different representations that encode content, emotion, and other factors of variations. Based on the input audio, a sequence of content representations are sampled from the distribution, which along with input face images are fed to a video generator to animate the face.

The researchers sourced three data sets to train and test the VAE: GRID, an audiovisual corpus containing 1,000 recordings each from 34 talkers; CREMA-D, which consists of 7,442 clips from 91 ethnically diverse actors; and LRS3, a database of over 100,000 spoken sentences from TED videos. They fed GRID and CREMA-D to the model to teach it disentangled phonetic and emotional representations, and then they evaluated the quality of generated videos using a pair of quantitative metrics, peak signal-to-noise ratio (PSNR) and Structural Similarity Index (SSIM).

The team says that their approach is on par, in terms of performance, on all metrics with other methods for clean, neutral spoken utterances. Moreover, they note that it’s able to perform consistently over the entire emotional spectrum, and that it’s compatible with all current state-of-the-art approaches for talking head generation.

“Our approach to variation-specific learnable priors is extensible to other speech factors such as identity and gender which can be explored as part of future work,” wrote the coauthors. “We validate our model by testing on noisy and emotional audio samples, and show that our approach significantly outperforms the current state-of-the-art in the presence of such audio variations.”

Let’s block ads! (Why?)

Big Data – VentureBeat

AI generates melodies from lyrics

September 1, 2019 Big Data

Generating sequences of musical notes from lyrics might sound like the stuff of science fiction, but thanks to AI, it might someday become as commonplace as internet radio. In a paper published on the preprint server Arxiv.org (“Conditional LSTM-GAN for Melody Generation from Lyrics“), researchers from the National Institute of Informatics in Tokyo describe a machine learning system that’s able to generate “lyrics-conditioned” melodies from learned relationships between syllables and notes.

“Melody generation from lyrics has been a challenging research issue in the field of artificial intelligence and music, which enables to learn and discover latent relationship between interesting lyrics and accompanying melody,” wrote the paper’s coauthors. “With the development of available lyrics and melody dataset and [AI], musical knowledge mining between lyrics and melody has gradually become possible.”

As the researchers explain, notes have two musical attributes: pitch and duration. Pitches are perceptual properties of sounds that organize music by highness or lowness on a frequency-related scale, while duration represents the length of time that a pitch or tone is sounded. Syllables align with melodies in the MIDI files of music tracks; the columns within said files represent one syllable with its corresponding note, note duration, and rest.

The researchers’ AI system made use of the alignment data with a long-short-term memory (LSTM) network, a type of recurrent neural network capable of learning long-term dependencies, with a generative adversarial network (GAN), a two-part neural network consisting of generators that produce samples and discriminators that attempt to distinguish between the generated samples and real-world samples. The LSTM was trained to learn a joint embedding (mathematical representation) at the syllable and word levels to capture the synaptic structures of lyrics, while the GAN learned over time to predict melody when given lyrics while accounting for the relationship between lyrics and melody.

To train it, the team compiled a data set consisting of 12,197 MIDI files, each paired with lyrics and melody alignment — 7,998 files from the open source LMD-full MIDI Dataset and 4,199 from a Reddit MIDI dataset — which they cut down to 20-note sequences. They took 20,934 unique syllables and 20,268 unique words from the LMD-full MIDI, and extracted the beats-per-minute (BPM) value for each MIDI file, after which they calculated note durations and rest durations.

Here’s one generated sample:

And here’s another:

After splitting the corpus into training, validation, and testing sets and feeding them into the model, the coauthors conducted a series of tests to determine how well it predicted melodies sequentially aligned with the lyrics, MIDI numbers, note duration, and rest duration. They report that their AI system not only outperformed a baseline model “in every respect,” but that it approximated well to the distribution of human-composed music. In a subjective evaluation during which volunteers were asked to rate the quality of 12 20-second melodies generated using the baseline method, the AI model, and ground truth, scores given to melodies generated by the proposed model were closer to those composed by humans than the baseline.

The researchers leave to future work synthesizing melodies with sketches of uncompleted lyrics and predicting lyrics when given melodies as a condition.

“Melody generation from lyrics in music and AI is still unexplored well [sic],” wrote the researchers. “Making use of deep learning techniques for melody generation is a very interesting research area, with the aim of understanding music creative activities of human.”

AI might soon become an invaluable tool in musicians’ compositional arsenals, if recent developments are any indication. In July, Montreal-based startup Landr raised $ 26 million for a product that analyzes musical styles to create bespoke sets of audio processors, while OpenAI and Google earlier this year debuted online creation tools that tap music-generating algorithms. More recently, researchers at Sony investigated a machine learning model for conditional kick-drum track generation.

Let’s block ads! (Why?)

Big Data – VentureBeat

AI generates interesting story endings

July 22, 2019 Big Data

Sophisticated natural language processing systems like OpenAI’s GPT-2 can craft speech that’s impressively humanlike, but those same AI often struggle with cogencY and coherency. In particular, they don’t pen compelling conclusions — AI-generated story endings tend to be generic and lacking in context.

This shortcoming motivated scientists at Carnegie Mellon University’s School of Computer Science to devise a method that creates more “diverse” endings for a given story. The key, they say, was training models to focus attention on important phrases of the story and promoting the generation of non-generic words.

“A story context is a sequence of sentences connecting characters and events. This task is challenging as it requires modeling the characters, events, and objects in the context, and then generating a coherent and sensible ending based on them. Generalizing the semantics of the events and entities and their relationships across stories is a non-trivial task,” wrote the coauthors. “We show that the combination of the two leads to more diverse and interesting endings.”

Above: A few of the proposed model’s outputs.

The team tapped seq2seq — a type of long short-term memory recurrent neural network architecture that’s capable of learning dependencies — with attention to create mathematical representations of words belonging to the context of the target story, and to learn those words’ relationships and translate them back into human-readable text. To incorporate key phrases from the story context, the researchers used an algorithm — RAKE — that picked out phrases and assigned them scores based on word frequency and co-occurrence, and then they manually sorted the phrases by their corresponding scores and discarded those below a certain threshold.

To generate endings, the scientists trained their model on the ROCStories corpus, which contains over 50,000 five-sentence stories. And to evaluate the model, they used DIST (Distinct), which calculates numbers of distinct unigrams (the contiguous sequence of n items from a given sample of text or speech), bigrams (a pair of consecutive written units such as letters, syllables, or words), and trigrams (a trio of consecutively written units) in the generated responses divided by the total numbers of unigrams, bigrams, and trigrams.

In a separate test, they trained Google’s BERT on the open source Story-Cloze task to compare their model with a baseline by selecting the correct ending of a story given two choices.

So how’d the AI perform? Let’s just say a Pulitzer isn’t in the cards. While it was the top performer in DIST and managed to get a Story-Cloze test accuracy of 72%, it occasionally generated nonsensical endings like “katie was devastated by himself and dumped her boyfriend” or referred to nouns with incorrect pronouns (“katie,” “himself”).

The researchers concede that further work is needed to ensure the outputs “entail the story context at both semantic and token level,” and that they’re logically sound and consistent. Still, they assert that they’ve “quantitatively” and “qualitatively” shown that their model can achieve “meaningful” improvements over the baselines.

Let’s block ads! (Why?)

Big Data – VentureBeat

Tag Archives: generates

LinkedIn’s AI generates candidate screening questions from job postings

IBM’s AI generates new footage from video stills

Microsoft’s AI generates 3D objects from 2D images

Tips for Creating a Landing Page That Generates Conversions

1. Make Sure Your Headline Is Clear and Catchy

2. Keep Your Copy Brief and to the Point

3. Always Include Important Information

4. Use Design Best Practices to Make Your Landing Page Visually Appealing

5. Make It Easy for Individuals to Sign Up

6. Include a Clear CTA That Is Easy to Find

7. Feature Key Customer Logos and Reviews

8. A/B Test Elements to Optimize Your Efforts

Act-On Can Help You Build Engaging Landing Pages That Spark the Customer Journey

Related Blog Posts

MIT CSAIL’s TextFooler generates adversarial text to strengthen natural language models

IBM’s biology-inspired AI generates hash codes faster than classical approaches

IBM’s Lambada AI generates training data for text classifiers

Microsoft’s AI generates high-quality talking heads from audio

AI generates melodies from lyrics

AI generates interesting story endings

Recent Posts

Categories

Archives