GCRI · 量子技術
量子コンピューティング
Quantum computers use qubits that can be in superpositions and become entangled, which in principle lets some problems be solved far faster than on classical computers. Current machines are noisy and small, and useful fault-tolerant computing requires error correction that is only beginning to be demonstrated. This page sets out how quantum computers work and fail, how progress is measured honestly, and what evidence serves the institutions deciding how to engage.
Quantum computing on the record.
Quantum computing: how the risk develops.
Hardware, algorithms and actual maturity. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.
Qubits are fragile
Qubits lose their quantum state through interaction with their environment. Gate errors and decoherence limit how long a computation can run before results become noise.
Error correction needs overhead
Fault-tolerant computing encodes each logical qubit in many physical qubits and corrects errors continuously. It only helps once physical error rates fall below a threshold.
Advantage is problem-specific
Quantum algorithms speed up particular problems, such as simulating quantum システム or factoring. For many tasks classical computers remain as good or better.
Claims need classical comparison
Demonstrations of quantum advantage are often followed by improved classical methods. Honest comparison requires the best classical approach, not a convenient one.
Quantum testbed.
このテーマを追跡するためにチームが使う画面：その指標、状態、決定。名前は例であり、値はイメージです。
Quantum testbed
Quantum computing · Testbed lab · Device QD-28
量子ビット
| Q0 | 121 | 49 | 0.03 | Drifting |
| Q1 | 118 | 24 | 0.03 | Drifting |
| Q2 | 94 | 41 | 0.03 | Drifting |
| Q3 | 73 | 93 | 0.05 | Drifting |
| Q4 | 50 | 48 | 0.02 | Drifting |
Device
- Topology
- heavy-hex, illustrative
- 2Q fidelity
- 0.994
- アクセス
- by committee
- ラベル
- 観測
Quantum computing matters to the domains through simulation and through cryptography.
- エネルギー結合Simulation of battery and catalyst chemistry is a proposed early use.What to watchDemonstrated against classical baselines
- 健康結合Future cryptanalysis threatens long-lived health data.What to watchMigration of health systems
- 食料結合Fertiliser catalysis is often cited as a target problem.What to watchClaims separated from results
Status reviews are designed for the Exponential Technology series, separating demonstrations from roadmaps. No date for a 暗号解読に関係する量子コンピュータ の実現時期を既知として示すことはありません。
- Peer-reviewed results with classical comparisons.
- Corrections to over-stated claims.
How quantum computing is modelled, measured and tested.
モデル、その仮定、データ、不確実性を、科学者やエンジニアが確認する形で示します。構造は公表された手法に従っています。描かれた値は例示であり、知見ではありません。
システム・モデル
The couplings between the systems in quantum computing: signs of influence, a reinforcing and a balancing loop, and cross-links where stress jumps.
何を、どの程度うまく測定しているか。
各指標には、その単位、更新頻度、背後にある出典の種類、値が持つことになるエビデンスのラベルが付いています。値は表示しません：プログラムが自らの権限のもとで値を提供します。
| 指標 | 単位 | 周期 | 出典の種類 | ラベル |
|---|---|---|---|---|
| Physical qubit count | qubits | per announcement | Published device results | 報告値 |
| Two-qubit gate fidelity | percent | per device | Benchmark results | 報告値 |
| Logical qubit error rate | errors per logical operation | per experiment | Peer-reviewed experiments | 観測 |
| Coherence time | microseconds | per device | Device characterisation | 報告値 |
| Quantum volume or similar holistic benchmark | benchmark score | per device | Published benchmarks | 報告値 |
| Demonstrations surviving classical challenge | 件数 | レビューごと | Literature review | 推定 |
エビデンスの出所。
名前の明示された公開データセットと機関の記録。それぞれにライセンス区分があります。合意に基づく出典は、その条件のもとでのみ使用し、再公開することはありません。
| 出典 | 提供するもの | ライセンス区分 |
|---|---|---|
| arXiv quant-ph and peer-reviewed journals | Primary experimental and theoretical results | Open access preprints; journal licence per article |
| Metriq benchmark platform (Unitary Foundation) | Community-submitted quantum benchmark results | Open, licence to check |
| Vendor technical documentation and roadmaps | Device specifications and stated targets | Public documents, terms per vendor |
| National academies and agency assessments of quantum computing | Independent assessments of progress and prospects | Public, terms per publisher |
| QED-C reports | Industry consortium analyses of use cases and benchmarks | Public reports, terms to check |
誰が決め、エビデンスが何に役立つか。
Quantum Technologies prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.
| 誰が決定するか | 質問 | このページが提供するエビデンス |
|---|---|---|
| Research funders and science ministries | How much to invest in quantum computing research and access | Progress summaries; independent assessments |
| Organisations exploring quantum use | Whether a problem is a plausible quantum use case | Use case assessments with classical baselines |
| Security agencies | How progress affects cryptographic migration timelines | Error correction and scale evidence |
このトピックが生み出す成果物
- Progress summaries that separate peer-reviewed results from roadmaps
- Benchmark interpretation guides
- Use case assessments with classical baselines
- Evidence notes for research funders
専門家が問うこと。
The central question is what has been demonstrated, not what has been promised.
- How many qubits does a device have, and how good are they: what are their error rates and coherence times?
- Can errors be corrected faster than they accumulate, and at what cost in physical qubits per logical qubit?
- Which algorithms offer a demonstrated advantage over the best classical methods, and for which problems?
- Are claims of quantum advantage robust to improved classical algorithms?
- Which hardware approaches are likely to scale, and what limits them?
メカニズムと要因。
Quantum computers gain their power from interference: algorithms are designed so that wrong answers cancel and right answers reinforce. That power is fragile. Qubits lose their quantum state through interaction with the environment, and every operation introduces some error. Quantum error correction encodes one reliable logical qubit in many physical qubits, but only works if physical error rates are below a threshold, and it multiplies the hardware required.
Known algorithms with large demonstrated speed-ups are few. Factoring and discrete logarithms, which underpin public-key cryptography, are the best-known examples; simulation of quantum システム in chemistry and physics is another. For many optimisation and machine-learning problems, a practical advantage has not been established, and improved classical algorithms have repeatedly narrowed claimed gaps.
どのように研究され、手法がどこで止まるか。
| 手法またはデータ | 表示内容 | 限界 |
|---|---|---|
| Benchmarking of gate fidelity and coherence | Quality of physical qubits | Benchmarks can be chosen to flatter a device |
| Error-correction experiments | Whether logical qubits outperform physical ones | Early demonstrations are small and short-lived |
| Algorithm complexity analysis | Theoretical speed-up and resource estimates | Resource estimates depend on hardware assumptions |
| Classical simulation and verification | Whether results could be reproduced classically | Verification becomes harder as devices grow |
| Peer-reviewed literature versus announcements | What has been independently assessed | Announcements often precede, or replace, peer review |
Where this page’s references lead.
このページの読解の基準となる機関と文書。状態と版は変わります。各リンクは発行機関へ移動します。
質問と回答。
量子ビットとは何ですか？
What is quantum advantage?
What is fault-tolerant quantum computing?
Why does qubit count alone mislead?
Does the platform predict when large quantum computers will arrive?
このページからエビデンスを提供する。
報告書 is the Consortium’s quarterly publication. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work.
- メールゲートウェイ GW-GCRIこのページから送信するエビデンス、訂正、事例、手法、レビュー、翻訳、プロファイルへの入力。このページのタグが付いて届くため、担当の編集者が受け取ります。
- Zenodoオープンな寄託寄託してコミュニティに提出するORCID iD を使って Zenodo に自分の記録を公開し、そこから Reports コミュニティに提出してください。キュレーターが審査し、DOI は引き続きあなたのものです。
- 報告書GCRI貢献のすべての方法を見る貢献の種類、それぞれの審査方法、受け取るクレジット、ライセンス、決して受け付けないもの。
個人データ、重要インフラの機微な詳細、共有する権利のないものは決して送らないでください。寄稿は利用前にレビューされます。送付によって著作者としての権利や編集上の権利が生じることはなく、スポンサーシップで調査結果を買うことは決してできません。
手法、データセット、または問いを持ち込んでください。
情報提供と問い合わせは本日から受け付けています。問い合わせには参照番号が付き、次の手順が示されます。申請、支払い、計算資源の割り当て、登録は、担当チーム、規約、審査体制が整ってから開始します。
法的文書。
このサイトの地図、データ、図表は、背景と一般的な情報として示すものです。助言、予報、警報ではありません。情報の限界は免責事項とデータに関する注意事項に記載しています。