Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesSome links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Yes—but in a specific, limited sense. The 2018 study behind this claim trained a deep neural network to recognize which chemical element belongs at a crystallographic site by reading that site’s local structural environment. It then used the learned relationships to rank element combinations that might fit known crystal-structure templates. This is computational candidate ranking, not an autonomous laboratory discovery: the reported work did not synthesize or experimentally confirm a new material.
The approach matters because the number of possible compounds grows rapidly when a material contains three or more elements. A model that narrows that search can help researchers decide which compositions deserve synthesis attempts first.
As an Amazon Associate I earn from qualifying purchases.
What did the AI actually learn?
The model learned a link between a site’s structural surroundings and the chemical element occupying that site. In a crystal, each atom sits in a recurring coordination environment: neighboring atoms, bond directions, distances and the topology of the surrounding polyhedron. The study encoded those features as a multiperspective atomic fingerprint, rather than as a photograph or a conventional image of a specimen.
That distinction is important. The network was not looking at a microscope picture and naming a substance. It received numerical descriptors of crystallographic sites and learned statistical patterns connecting local geometry with element identity. Once trained, it could ask whether a proposed element was plausible for a particular site in a known structural arrangement.
#1 Best Overall
- Read Before You Buy — No Video Output: These adapters support charging and USB 2.0 data transfer, but cannot transmit video signals. Except for standard USB webcams (which use USB data only), they are not compatible with HDMI/DisplayPort cables, video-capable USB-C hubs, or docking stations with video output.
- Convert USB-A Ports to USB-C: Designed to connect USB-C earphones, cables, flash drives, card readers, and other USB-C accessories to standard USB-A ports. Plug-and-play with no drivers or software required.
- Aluminum Alloy Housing: Built with a sturdy aluminum alloy shell that aids in heat dissipation and protects against daily wear and scratches. Designed to maintain a stable and secure connection.
- Compact & Travel-Friendly: The ultra-compact design allows the adapter to stay plugged into your device without blocking adjacent ports or adding bulk, reducing wear and tear on your original USB ports.
- 12-Month Warranty: Backed by a 12-month manufacturer warranty for peace of mind. Designed to meet strict quality control standards for reliable everyday performance.
The method therefore addresses a constrained materials question: given a structure template and possible elements, which assignments look most compatible with patterns seen in existing crystal data?
How does the crystal-prediction workflow work?
The study’s process can be understood as a sequence from crystallographic data to a ranked list of compositions.
- Collect solved structures. The training material consisted of approximately 50,000 crystal structures. Each structure supplied sites with known element identities and local environments.
- Convert each site into fingerprints. Multiperspective atomic fingerprints describe the coordination topology around a site from several structural perspectives. They turn geometric and connectivity information into features a neural network can process.
- Train the deep neural network. During training, the network adjusted its internal parameters to associate fingerprint patterns with the elements known to occupy those sites.
- Apply the model to a structure template. The researchers used templates derived from known binary and ternary structures. The framework was therefore anchored to an existing arrangement of sites, rather than asked to invent every possible crystal lattice from nothing.
- Rank candidate compositions. For a template, the model scored possible element assignments and returned an ordered list. A high ranking means the assignment resembles combinations supported by the learned data; it does not prove that the compound is stable or synthesizable.
This ranking is useful because a chemist can focus experimental planning on a manageable subset of possibilities instead of treating every chemically conceivable combination equally.
What kinds of materials did it predict?
The prediction task focused on compositions that could occupy known structural templates. The study considered binary structures, containing two elements, and ternary structures, containing three. It was not an unrestricted generator of arbitrary crystal structures, phases or processing conditions.
| Aspect | What the study covered | What it did not establish |
|---|---|---|
| Input representation | Multiperspective fingerprints of local crystallographic environments | Images, spectra or a complete description of synthesis conditions |
| Structural scope | Templates derived from known binary and ternary crystal structures | Every possible lattice type or unconstrained structure generation |
| Output | A ranked list of likely element compositions for a template | A guaranteed formula, phase diagram or processing recipe |
| Scientific status | Computational candidates for follow-up | Experimental synthesis and confirmation by this model |
In practical terms, the network could suggest that certain elements are more plausible occupants of the sites in a familiar arrangement. Researchers would still need to determine whether the proposed composition forms the intended phase, under what conditions, and with which measurable properties.
Rank #2
- 5-in-1 USB-C Hub: Experience comprehensive connectivity featuring a Power Delivery input, two USB-A 2.0 ports, a USB-A 3.0 port, and an HDMI port. (Note: The USB-C power delivery input port is only for connecting an external wall charger to power your laptop and cannot power peripheral devices.)
- 90W Pass-Through Charging: Achieve optimal charging with 90W pass-through power to your laptop, supported by a total input of 100W, with the hub reserving 10W for operational efficiency. (Note: Wall charger not included.)
- Quick Data Transfers: Accelerate your productivity with rapid data transfers using a high-speed 5Gbps USB 3.0 port and two 480Mbps USB 2.0 ports.
- 4K HDMI Display: Enhance your visual experience with a hub capable of delivering 4K resolution at 30Hz in both mirror and extend modes. Please note that this hub is compatible with MacBook (macOS 12 and newer), Windows 10 and 11, ChromeOS, and laptops equipped with DP Alt Mode and Power Delivery. Note: This device is not compatible with Linux.
- What You Get: Anker USB-C Hub (5-in-1, 4K HDMI), welcome guide, 18-month warranty, and our friendly customer service.
How large was the dataset, and how well did the model perform?
Kevin Ryan, Jeff Lengyel and Michael Shatruk reported results from a dataset of approximately 50,000 crystal structures. On unseen test structures, the known composition appeared among the model’s ten highest-ranked candidates in about 30% of cases. The study’s abstract states: "In ~30% of cases, the known compositions were found among top-10 most likely candidates proposed by the model."
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minute| Reported item | Meaning |
|---|---|
| Data volume | Approximately 50,000 crystal structures used by the study |
| Evaluation set | Unseen test structures, rather than only structures used for training |
| Result | About 30% of known compositions appeared in the top ten ranked candidates |
| Interpretation | A retrieval result for this study’s test setup, not a universal accuracy rate or proof of synthesis |
The figure tells you how often the correct known composition was near the top of the list under that evaluation. It does not say that 30% of all undiscovered materials will work, that every tenth suggestion is experimentally viable, or that the network has general knowledge of chemistry outside the represented data.
Can AI find new materials without experiments?
No. A ranked composition is a hypothesis, not a material in a vial. The available account of this study establishes computational predictions and a proposed role in guiding synthetic work; it does not establish that the model itself made or verified a new compound in the laboratory.
Experimental confirmation requires a synthesis route and measurements that distinguish the intended phase from impurities, competing phases or an amorphous product. Depending on the research question, scientists may need diffraction, composition analysis and property measurements. None of those steps is supplied by the neural network’s ranking alone.
The most accurate description is therefore AI-assisted candidate selection. The model can reduce the number of compositions that deserve attention, while chemists and materials scientists decide what to make and test.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Rank #3
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
Why are ternary and more complex materials a difficult target?
Adding an element expands the combinatorial search space. A binary system already offers many element pairs and possible arrangements; a ternary system adds another chemical degree of freedom, along with more possible site occupations, oxidation-state combinations and competing phases. The 22 June 2018 report in Chemistry World highlighted this growth in possibilities and described candidate ranking as a way to guide synthesis in more complex systems.
Known templates provide a useful boundary for the problem. Instead of searching every imaginable structure and composition simultaneously, the researcher starts with an arrangement that has already been observed and asks which element combinations may fit it. That restriction makes the task tractable, but it also limits what the model can propose: a genuinely new structure outside the template library may never appear in the candidate list.
How should the 30% top-ten result be interpreted?
It is best read as a retrieval statistic tied to a particular dataset, model and test protocol. The result says that, in roughly three out of ten unseen test cases, the known answer was found within the first ten suggestions. It does not measure the fraction of successful laboratory syntheses, nor does it provide a probability that an individual candidate will form.
Two additional cautions follow. First, the test structures and their chemical distribution determine how difficult the ranking task is; performance on a different collection could change. Second, a candidate can be chemically plausible yet fail because of thermodynamic competition, kinetic barriers, processing constraints or an unrecognized structural change.
A 2019 review, Recent advances and applications of machine learning in solid-state materials science, also warns generally that leave-one-out cross-validation is not a sound way to measure extrapolation ability. That methodological warning should not be attributed to the Ryan study without evidence that the study used that validation design. It does reinforce a broader point: a convincing evaluation must test how a model behaves on genuinely unfamiliar chemistry, not merely on near neighbors of its training examples.
Rank #4
- Dual Converters, Infinite Potential：Includes 2× USB C male to USB A female adapters and 2× USB A male to USB C female adapters. Perfect for a wide range of uses—tablets with Bluetooth keyboards, expand USB ports on macbook, and more. Two different converters for all your daily needs
- Next-Level 10Gbps & 3A Charging: No more slow 480Mbps, this usb to usb c adapter has a transfer speed of up to 10Gbps, allowing you to do more transferring in less time. This usb adapter fits both USB A and USB C charger, supporting up to 3A fast charging
- Upgraded Exquisite Craftsmanship: With an aluminum alloy housing and metal connector, the usbc to usb adapter is extremely durable and sturdy. Rigorously tested to withstand more than 10,000 times of plugging and unplugging, ensuring long-lasting performance
- Broad Compatible: The usb c to usb adapter widely supports all USB C/ USB A devices like laptops, tablets, cellphones, car chargers, and phone chargers. Such as compatible with MacBook Pro/Air 2023/2022, Thunderbolt 4/3 Devices,Apple MagSafe Watch 9/8/7/SE/Ultra, iPad Pro 2022/2021, Samsung Galaxy S23/S20/S10, and iPhone 17/16/15 Pro. Plug and play
- Please Note: To reach 10Gbps speed, keep the cable under 3.3 ft. For USB A Male to USB C adapters, try flipping the USB C connector. USB C Male to USB A adapters support bidirectional 10Gbps transfer within 3.3 ft
What is the difference between this approach and broader materials AI?
Materials machine learning covers several distinct tasks. Comparing the task and evidence prevents a site-classification model from being mistaken for a universal crystal generator.
| Approach | Typical input | Typical output | Main question |
|---|---|---|---|
| Site identity from fingerprints | Local coordination topology encoded as atomic fingerprints | Likely element assignments or compositions in a template | Which elements fit these known environments? |
| Structure classification from diffraction data | Diffraction patterns or related measurements | Phase or structure labels | What structure is present in a measured sample? |
| Property prediction | Composition, structure or derived descriptors | Estimated property such as a transition temperature or band gap | What might this material do? |
| Structure generation | Chemical constraints, compositions or latent representations | Candidate lattices and atomic arrangements | What structures could exist? |
| Experiment-guided discovery | Model rankings combined with synthesis and measurements | Validated compounds and updated models | Which candidates survive laboratory testing? |
The 2018 work belongs to the first category, with a composition-ranking use in known templates. A broader generator may explore structures the template method cannot represent, while an experimental campaign supplies evidence that a computational score alone cannot.
What would a researcher do with the ranked candidates?
A sensible follow-up turns the model’s list into a sequence of chemical and experimental checks.
- Inspect the structural fit. Confirm that the proposed elements can occupy the template’s sites without violating the intended coordination geometry.
- Check chemical plausibility. Consider charge balance, likely oxidation states and known competing compounds. A high model score does not replace this chemical reasoning.
- Prioritize feasible synthesis. Select candidates for which precursors, temperatures, atmospheres and handling methods are available.
- Make and characterize samples. Use appropriate measurements to determine whether the target phase formed and whether unwanted phases are present.
- Feed results back into discovery. Successful and failed attempts can become new training data, provided they are represented consistently and evaluated without leakage.
This workflow treats AI as a triage tool. Its value is greatest when the unranked search space is too large for a laboratory to explore exhaustively, but the final decision still depends on domain expertise and measurement.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What are the main limitations and failure modes?
- Template dependence: candidates are tied to known binary and ternary structural arrangements, so unfamiliar structure types may be missed.
- Data bias: a database reflects what has already been solved and reported. Common, stable or easily measured compounds can dominate the learned patterns.
- Local-descriptor limits: a fingerprint of coordination topology may not encode every long-range interaction, defect, disorder effect or synthesis-dependent phase.
- Ranking is not feasibility: the model can place a composition highly even when kinetics, competing phases or processing conditions prevent its formation.
- Evaluation leakage risk: closely related structures across training and test sets can make a task look easier than extrapolating to genuinely new chemistry.
- No property guarantee: identifying a plausible composition says nothing by itself about conductivity, magnetism, strength, toxicity or other performance measures.
These failure modes do not make the method useless. They define the question it can answer reliably enough to be valuable: which candidates deserve attention first, given the structural assumptions and data used to train the model?
Best Value
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
So, can AI find new materials?
AI can help researchers find candidates for new materials by learning from existing crystal structures and ranking compositions that fit known templates. In the Ryan, Lengyel and Shatruk study, that capability was demonstrated computationally on unseen test structures, with the known composition appearing in the top ten in about 30% of cases.
Calling the system an autonomous discoverer would overstate the evidence. The study did not show autonomous synthesis or experimental confirmation. Its realistic contribution is a bridge between a large crystallographic search space and a shorter, more informed list of experiments.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Frequently Asked Questions
What can AI predict about a crystal?
A model of this kind can estimate which elements are plausible at sites within a specified structural template and rank complete compositions accordingly. It does not automatically determine every property, the best synthesis recipe or the final phase obtained in the laboratory.
Does a high-ranked composition mean the compound is stable?
No. Stability depends on thermodynamics, competing phases, temperature, pressure and kinetics, none of which is guaranteed by a site-fingerprint score. Stability calculations and experiments are needed to test that question.
Why use known crystal templates instead of generating structures from scratch?
Templates constrain the search to arrangements that have already been observed, making candidate ranking more manageable. The trade-off is that a material requiring a wholly new structure type may be excluded from the search.
Who conducted the 2018 study?
Kevin Ryan, Jeff Lengyel and Michael Shatruk published the work, titled "Crystal Structure Prediction via Deep Learning," in the Journal of the American Chemical Society in 2018. A contemporaneous report appeared in Chemistry World on 22 June 2018.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
