Autonomy, motori di ricerca per archivi sonori

Un altro comunicato stampa che mi ha incuriosito in questi giorni viene da Autonomy, una societa britannica fondata da un gruppo di matematci di Cambridge ed è specializzata nei motori di ricerca semantici. Una ricerca basata cioè su algoritmi di tipo non booleaiano ma baynesiano, capaci di affrontare il contenuto in chiave analogica, analizzandone il senso. Il nuovo prodotto di Autonomy è rivolto in particolare ai contenuti audio e alle applicazioni in cui il computer deve analizzare materiali audio di archivio, riconoscere l'identità dei parlanti e individuare termini e contesti precisi. Sembra che questo tipo di applicazioni sia sempre più richiesto tra le grandi aziende quotate americane, perché le autorità di Borsa impongono a queste aziende di archiviare non solo i messaggi di mail ma anche le telefonate e le conversazioni. E' sugli archivi audio che la piattaforma Autonomy eDiscovery è in grado di lavorare, estraendo appunto anche da centinaia di ore di conversazione determinati concetti espressi da chi parla e il timbro di voce dei singoli individui.
Il campo di applicazione delle soluzioni Autonomy è tipicamente aziendale, ma la possibilità di effettuare una ricerca diretta in un archivio di informazioni non strutturate come può essere un deposito multimediale di contenuti audio e video (alla AT&T, per esempio, lavorano su sistemi in grado di classificare automaticamente i notiziari televisivi, in base al riconoscimento di personaggi e situazioni) apre la strada a nuove opportunità di fruizione dei contenuti archiviati dagli enti radiotelevisivi. Quanto possano essere efficaci questi strumenti in un contesto di larga scala come potrebbe essere il sito Web di una emittente che funga da interfaccia a un archivio storico digitalizzato è ancora difficile dirlo. Anche perché non sono poi così numerosi gli esempi di archivi digitali estesi e consolidati. In molte situazioni i materiali sonori della radio - tanto per focalizzarci sul nostro orticello - sono per evidenti questioni storiche tuttora custoditi su supporti analogici, mentre cominciano a fiorire i progetti di digitalizzazione di archivi multimediali pubblici o privati (organizzazioni tipo Istituto Luce o le varie fonoteche nazionali, tanto per intendersi). Siamo ancora all'inizio ma strumenti software come questo di Autonomy giocheranno sicuramente un ruolo

Autonomy unveils most advanced audio EDISCOVERY for end-to-end ediscovery solution

Reduces Volume of Audio Recordings Requiring Review by 95 Percent; Saving Costs and Improving Investigators’ Confidence

Cambridge, UK and San Francisco, Calif. – Feb. 2, 2009 – Autonomy Corporation plc (LSE: AU. or AU.L), a global leader in infrastructure software for the enterprise, today unveiled the world’s most advanced audio eDiscovery solution that is integrated into the company’s end-to-end eDiscovery solution. Users of the Autonomy’s solution have reported a 95 percent reduction in the volume of audio recordings requiring review, significantly reducing eDiscovery costs while simultaneously boosting investigators’ confidence. This is achieved due to Autonomy’s unique ability to automatically understand the concepts and identify speakers in hundreds of hours of audio recordings, as well as its ability to link audio with other data types, including email.
Today, audio has become one of the most common forms of evidence requested by the United States Securities and Exchange Commission (SEC), with legacy audio processing costing organizations 80 percent more per gigabyte than email. Recent regulations, such as the SEC 17a (4) in the United States and the FSA COBS 11.8 in the United Kingdom requiring that all content formats, including audio, be captured and maintained for compliance purposes are prompting global organizations to embrace Autonomy’s Meaning-Based Computing for an end-to-end eDiscovery process.
Unlike legacy eDiscovery methods for audio, which are difficult to use, time-consuming and costly with the results often failing to meet the court’s expectations, Autonomy fully integrates this traditionally untapped asset into all phases of the investigative process, saving costs and reducing risks of legal penalty for organizations. As the leading eDiscovery software & service provider, Autonomy utilizes a fundamentally different approach to eDiscovery and legal review by simplifying, unifying and delivering high quality results across all content formats. Autonomy provides a single platform for understanding and processing all digital information. Only Autonomy can perform keyword and conceptual search, speech analytics, advanced visual analysis, redaction and all aspects of the legal review and production process on all electronic forms, formats and languages in a single unified environment. This results in significantly faster, more relevant and lower cost for eDiscovery.
“The current economic environment has brought eDiscovery to the forefront with investigators increasingly looking for potentially incriminating evidence in previously overlooked formats, such as audio and video,” commented Mike Sullivan, CEO of Autonomy ZANTAZ. “Autonomy’s Audio eDiscovery allows legal teams to quickly evaluate audio assets and prioritize review, resulting in substantial cost-savings and providing organizations with a distinct advantage in and out of court.”

Autonomy’s Audio eDiscovery solution offers unique capabilities, including:

• Adjustable search confidence – Autonomy’s Confidence Bar allows the investigator to easily expand or tighten the focus of each search, a clear advantage when seeking the unknown needle in a haystack or following a known data trail.
• Advanced audio redaction with redaction type and annotation – This Autonomy exclusive ability for legal teams to perform their own audio redactions in a manner virtually identical to text redactions will save countless hours and reduce costs significantly
• Speaker identification – Eliminates the costly and protracted process of contracting lawyers to listen to hundreds of hours of audio to identity files containing specific speakers.
• EDD capabilities for all major call center systems – Many call recording system utilize proprietary formats requiring a pre-processing step to reassemble the audio and associated meta and TAPI data to the original state. The Autonomy EDD process automates this step and provides complete documentation of the process.
• Message Tracer, Clustering, and Find Similar – Help investigators put audio into context with other content, connecting custodians and improving understanding of the flow of information.
• Phonetic, word spotting & conceptual searching - By automatically selecting the appropriate search technology for each task, the need for audio expertise is eliminated, search consistency is improved and the overall process is streamlined, reducing resource requirements and costs.

Please visit www.autonomy.com/ediscovery, for more information.

About Autonomy
Autonomy Corporation plc, a global leader in infrastructure software for the enterprise, is spearheading the meaning-based computing movement and was recently ranked by IDC as the clear market share leader in enterprise search with market share nearly double that of its nearest competitor. Autonomy's technology allows computers to harness the full richness of human information, forming a conceptual and contextual understanding of any piece of electronic data including unstructured information, be it text, email, voice or video. Autonomy's software powers the full spectrum of mission-critical enterprise applications including information access technology, pan-enterprise search, information governance, end-to-end eDiscovery and archiving, records management, business process management, customer interaction solutions, and video and audio analysis, and is recognized by industry analysts as the clear leader in enterprise search.