2024 · paper
ViTAL: Vision TrAnsformer based Low coverage SARS-CoV-2 lineage assignment
The group publication list records a Bioinformatics paper applying a vision-transformer approach to low-coverage viral lineage assignment.[105]
computer architecture, chip design and processing-in-memory
Leonid Yavits
לאוניד יביץ
Identity: verifiedThe Bar-Ilan Computer Architecture Development Group page identifies Leonid Yavits with the engineering department and matches the Hebrew roster transliteration. [104]
Documented foundation
Research & experience
Bar-Ilan computer-architecture faculty member with semiconductor entrepreneurship experience whose group develops chip architectures for AI and memory-centric computation.[104]
computer architecturecontent-addressable memoryprocessing-in-memorygenomics accelerators
CV and official profile
The group homepage provides a professional biography and publication navigation; no standalone current CV was verified.[104]
Selected work
Representative records, not a complete publication list. Metadata confirms attribution; it does not independently replicate a result.
Patent evidence
15 catalogued patent records · 14 identified families · family unassigned for 1 record
Coverage: Partial inventor search
Both pages of the public inventor index were inspected as discovery, followed by 15 individual primary patent-document reads, grouped into 14 additional families. These span Technion, VisionTech/Broadcom, Horizon Semiconductors and an unassigned individual application. The preserved US10878906B2 and its US20200051634A1 version were intentionally omitted from additions. Institutional profile reports 22 patent outputs, but its filtered inventory was inaccessible; neither that output count nor this subset is a worldwide family census.
Hybrid processor
US10996959B2 · Published 2021-05-04
Published patent document inspected
Publication assignee: Technion Research and Development Foundation Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[297][425]
Video encoding and video/audio/data multiplexing device
US9247263B2 · Published 2016-01-26
Published patent document inspected
Publication assignee: Broadcom Corp
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[337][425]
System and method for video and audio encoding on a single chip
US8270479B2 · Published 2012-09-18
Published patent document inspected
Publication assignee: Broadcom Corp
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[336][425]
Targeted content with broadcast material
US20090235312A1 · Published 2009-09-17
Published patent document inspected
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[315][425]
Automotive entertainment, communication, navigation and control center
US20090210155A1 · Published 2009-08-20
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[314][425]
Show 10 moreShow fewer patent records
Audio Processor
US20090055005A1 · Published 2009-02-26
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[313][425]
Set top box with transcoding capabilities
US20080263621A1 · Published 2008-10-23
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[312][425]
Hybrid hierarchical motion estimation for video streams
US20080260033A1 · Published 2008-10-23
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[311][425]
Media processor with an integrated TV receiver
US20080240230A1 · Published 2008-10-02
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[310][425]
Stream multiplexer/de-multiplexer
US20080240093A1 · Published 2008-10-02
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[309][425]
Method and apparatus for video pixel interpolation
US20080174694A1 · Published 2008-07-24
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[308][425]
Integrated circuit, an encoder/decoder architecture, and a method for processing a media stream
US20080120676A1 · Published 2008-05-22
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[307][425]
Home gateway for multiple units
US20080120675A1 · Published 2008-05-22
Published patent document inspected
Publication assignee: Horizon Semiconductors Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Assignee means the observed original-assignee field; no current ownership conclusion.[306][425]
Video encoding device
US6385244B1 · Published 2002-05-07
Published patent document inspected
Publication assignee: Visiontech Ltd
The inspected inventor field names Leonid Yavits. Rare full name, recurring Amir Morad co-inventor cluster, and documented VisionTech/Broadcom/Technion history support attribution across employer changes. Same inspected family: US6757329B2. One representative retained; variants intentionally grouped. Assignee means the observed original-assignee field; no current ownership conclusion.[332][333][425]
Resistive address decoder and virtually addressed memory
US10878906B2
Original report record
Preserved from the original report; see its cited evidence and limitations.[106]
Original report snapshot
Original evidence: verified record
Bar-Ilan's institutional record attributes the listed US patent to Leonid Yavits and co-inventors; current ownership and freedom-to-operate were not assessed.[106]
Records are counted separately from identified families. Author-reported entries are labelled and may still need publication verification. Inventorship, publication-time applicant and current ownership are different facts. No legal-status, patentability or freedom-to-operate conclusion is made.
Scores prioritize research fit from 1–10; they are not probabilities.
Review: Reviewed with limitations
Proposed capability matches, not confirmed relationships. Scores are analyst judgments with low forecast confidence; researcher interests, capacity and feasibility need confirmation.
Internal connections
14 candidates
Connection 1
Adam Teman
Original proposal
Proposal hypothesis: Teman's memory-circuit implementation and Yavits's processing-in-memory architecture connect directly at the energy cost of a genomic kernel.[1][104][105]
First test and score details
First test
Proposed first test: Replay one public sequence-classification workload through conventional and memory-centric models; compare memory accesses, latency and explicitly labelled energy estimates.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 4
Why this rank
Rank 1/14; fit 10/10 (4 topic overlap + 3 complementarity + 3 feasible first test). Preserved original co-membership proposal in o02. No strictly higher-scoring candidate displaces this original. Simulation is feasible without a tape-out; physical energy savings require later measured hardware. No automatic score boost for original membership. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 2
Alex Fish
Original proposal
Proposal hypothesis: Fish's low-power memory circuits complement Yavits's content-addressable genomics architectures at the circuit-to-workload boundary.[89][90][104][105]
First test and score details
First test
Proposed first test: Map the same public genome-classification kernel to two memory organizations and sweep precision and memory-error rate; report latency and energy proxies.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 4
Why this rank
Rank 2/14; fit 10/10 (4 topic overlap + 3 complementarity + 3 feasible first test). Preserved original co-membership proposal in o02. No strictly higher-scoring candidate displaces this original. Existing EnICS context supplies no score bonus; the score follows the concrete shared memory-workload question. No automatic score boost for original membership. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work. Original initiative conditions remain: o02 Alex Fish: Alternative or later circuit lead: assess memory/circuit precision and energy tradeoffs if the chosen implementation needs this expertise; security is outside the initial test.
Connection 3
Moti Medina
Proposal hypothesis: Medina's reliable hardware methods and Yavits's memory-centric architectures directly connect architecture choices with recoverable compute errors.[100][104][105]
First test and score details
First test
Proposed first test: Inject defined faults into a public genomic-kernel execution model and compare redundancy strategies on correct outputs and traffic overhead.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 4
Why this rank
Rank 3/14; fit 10/10 (4 topic overlap + 3 complementarity + 3 feasible first test). New pairing outside the frozen portfolio co-member graph. Physical error distributions and fabricated-hardware costs remain unverified. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Show 11 moreShow fewer internal connections
Connection 4
Tomer Kalisky
Original proposal
Proposal hypothesis: Kalisky can specify which cell-state distinctions must survive while Yavits changes the architecture executing a genomic classifier.[70][104][105]
First test and score details
First test
Proposed first test: Run an exact and approximate kernel on a public labelled single-cell dataset and compare rare-state retention alongside latency and memory traffic.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 3
Why this rank
Rank 4/14; fit 9/10 (3 topic overlap + 3 complementarity + 3 feasible first test). Preserved original co-membership proposal in o02. An added candidate, Moti Medina (10/10), ranks above this original because its stated pair-specific roles and first test score higher; this original is limited as follows: The selected single-cell task must actually map to the accelerator; pathogen-kernel performance cannot be transferred automatically. No automatic score boost for original membership. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 5
Orr Levy
Proposal hypothesis: Levy's multi-omics computations can supply a biologically meaningful workload for Yavits's memory-centric architecture.[5][6][104][105]
First test and score details
First test
Proposed first test: Profile a public biological-network kernel and compare memory-centric and conventional execution models on traffic, runtime and preserved rankings.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 3
Why this rank
Rank 5/14; fit 9/10 (3 topic overlap + 3 complementarity + 3 feasible first test). New pairing outside the frozen portfolio co-member graph. A kernel must be shown memory-bound before accelerator relevance is claimed. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 6
Shahar Alon
Proposal hypothesis: Alon's image-heavy in-situ sequencing could expose a memory-intensive kernel for Yavits's architecture research.[60][61][104][105]
First test and score details
First test
Proposed first test: Profile one open spot-matching or image-registration workload and compare an accelerated model with a CPU baseline on alignment error and memory traffic.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 3
Why this rank
Rank 6/14; fit 9/10 (3 topic overlap + 3 complementarity + 3 feasible first test). New pairing outside the frozen portfolio co-member graph. Genomics accelerator expertise does not establish that this imaging kernel maps efficiently. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 7
Gonen Singer
Proposal hypothesis: Singer's prescriptive resource allocation could choose how Yavits's memory-centric system spends a limited compute or energy budget.[79][80][104][105]
First test and score details
First test
Proposed first test: Replay a mixed synthetic kernel queue and compare fixed and adaptive placement on deadlines, traffic and energy proxies.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 2
Why this rank
Rank 7/14; fit 8/10 (2 topic overlap + 3 complementarity + 3 feasible first test). New pairing outside the frozen portfolio co-member graph. A meaningful system cost model is needed; energy proxies are not measured consumption. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 8
Izack Cohen
Proposal hypothesis: Izack Cohen's queueing and resource-allocation methods can evaluate contention in Yavits's memory-centric accelerators.[97][98][104][105]
First test and score details
First test
Proposed first test: Replay an open or synthetic workload trace through two memory-service policies and compare queue tails, throughput and traffic.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 2
Why this rank
Rank 8/14; fit 8/10 (2 topic overlap + 3 complementarity + 3 feasible first test). New pairing outside the frozen portfolio co-member graph. Service-operation expertise transfers as a modelling hypothesis, not demonstrated accelerator design. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 9
Yoli Shavit
Original proposal
Proposal hypothesis: Yavits can expose approximation and compression choices in a genomic accelerator and Shavit can propose generalization and uncertainty checks for the output.[104][105][127][128]
First test and score details
First test
Proposed first test: On a public genomic classification benchmark, sweep approximation settings and compare held-out calibration with classification fidelity and energy proxies.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 9/14; fit 7/10 (2 topic overlap + 3 complementarity + 2 feasible first test). Preserved original co-membership proposal in o02. An added candidate, Moti Medina (10/10), ranks above this original because its stated pair-specific roles and first test score higher; this original is limited as follows: Shavit's biological direction is documented as an announced programme, not an established experimental laboratory; the exact fidelity question comes first. No automatic score boost for original membership. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work. Yoli Shavit biological/multimodal programme evidence is a 2025 announcement; current laboratory operation and biological experimental capability are not established. Original initiative conditions remain: o02 Yoli Shavit: Conditional later role: test uncertainty and generalization only after a specific cross-domain fidelity question is preregistered.
Connection 10
Ephraim Zehavi
Proposal hypothesis: Zehavi's MIMO processing can provide a regular numerical kernel for Yavits's memory-centric architectures.[19][20][104][105]
First test and score details
First test
Proposed first test: Map a small detector kernel to conventional and memory-centric models and compare bit traffic, latency and decision errors.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 10/14; fit 7/10 (2 topic overlap + 3 complementarity + 2 feasible first test). New pairing outside the frozen portfolio co-member graph. A suitable detector must be selected; genomics acceleration is not proof of MIMO acceleration. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 11
Ran Gelles
Proposal hypothesis: Yavits's memory-centric execution could host a distributed task whose communication-state correctness Gelles specifies.[57][58][104][105]
First test and score details
First test
Proposed first test: Emulate a small distributed genomic-kernel exchange with message corruption and compare correct completions and memory traffic.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 11/14; fit 7/10 (2 topic overlap + 3 complementarity + 2 feasible first test). New pairing outside the frozen portfolio co-member graph. The workload must actually require interaction; adding a protocol to a local kernel has no demonstrated value. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 12
Shimrit Shtern
Identity check needed: Conditional proposal: confirm the researcher identity and research interests before assessing this match.
Proposal hypothesis: Conditional on confirmation, Shtern can formulate uncertain workload constraints for Yavits's memory-centric architecture choices.[83][84][85][104][105]
First test and score details
First test
Conditional proposed first test: After identity confirmation, compare nominal and robust memory-allocation decisions on a synthetic kernel mix using latency and energy proxies.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 12/14; fit 7/10 (2 topic overlap + 3 complementarity + 2 feasible first test). New pairing outside the frozen portfolio co-member graph. The match is a transferable optimization role, not evidence of chip-design expertise for Shtern. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work. Conditional: confirm Shimrit Shtern identity mapping and current institutional affiliation before any internal team assignment. The BIU directory and current Technion appointment leave affiliation unresolved. Documented optimization expertise is conditional on that mapping; excluded from confirmed-team claims.
Connection 13
Yatir Sadia
Proposal hypothesis: Sadia can characterize thermal material constraints and Yavits can expose memory-centric workload heat and power profiles.[31][32][104][105]
First test and score details
First test
Proposed first test: Compare two compute duty cycles using a simple thermal model; report peak temperature and lost throughput for an explicitly assumed material stack.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 1
Why this rank
Rank 13/14; fit 6/10 (1 topic overlap + 3 complementarity + 2 feasible first test). New pairing outside the frozen portfolio co-member graph. This is a thermal-computing feasibility question, not documented chip-cooling expertise for Sadia. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work.
Connection 14
Alexandra Simonovsky
Identity check needed: Conditional proposal: confirm the researcher identity and research interests before assessing this match.
Proposal hypothesis: Conditional on confirmation, Simonovsky could coordinate a reproducible prototype exercise using a small compute workload specified by Yavits.[16][17][104][105]
First test and score details
First test
Conditional proposed first test: After identity confirmation, prepare a software-only workload brief and common latency/memory-report template; pilot whether submissions are comparable.
Score components
- complementarity
- 2
- feasible first test
- 1
- topic overlap
- 1
Why this rank
Rank 14/14; fit 4/10 (1 topic overlap + 2 complementarity + 1 feasible first test). New pairing outside the frozen portfolio co-member graph. No access to hardware, research appointment or computer-architecture expertise is inferred for Simonovsky. Equal scores use existing-first, then stable researcher ID.
Conditions
Analyst proposal hypothesis, not an established collaboration, commitment, evidence-confidence rating or assessment of researcher quality; confirm participation and access before work. Conditional: confirm Alexandra Simonovsky identity and current role before any team assignment. Evidence supports AI-agent event coordination only; research appointment, teaching role and technical research expertise are not verified. Excluded from confirmed-team claims.
External connections
10 candidates
Connection 1
Onur Mutlu
ETH Zurich
Original proposal
Mutlu's official research page documents processing-in-memory and genome-analysis systems, creating a methodologically close but independently benchmarkable match; no willingness or current collaboration is asserted.[92][104][107][481]
First test and score details
First test
Implement a trace-driven prototype on public pathogen reads and measure energy proxy, latency and classification fidelity against CPU and GPU baselines.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 4
Why this rank
Rank 1/10 after semantic revision; analyst score 10 = max(1, 4+3+3): topic overlap 4/4, complementarity 3/3, feasible first test 3/3. A bounded offline comparison is specified; required datasets and domain assumptions must still be checked. Original remains first under these components; original status breaks equal-score ties only, without a prestige bonus.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified.
Connection 2
Lior S. Pachter
California Institute of Technology
Proposed capability match: Leonid Yavits's computer architecture, content-addressable memory can be paired with Lior S. Pachter's documented single-cell sequencing, RNA biology for energy-bounded pathogen surveillance accelerator. The specific contribution is genomic reference-pipeline reproducibility; this transfer is an analyst hypothesis.[104][483]
First test and score details
First test
Compare preprocessing and sequence-quantification variants before freezing the accelerator workload using public pathogen-read traces and a simulated associative-memory classification kernel. Compare classification disagreement and sensitivity to read preprocessing with CPU/GPU or conventional memory traces at matched classification fidelity.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 3
Why this rank
Rank 2/10 after semantic revision; analyst score 9 = max(1, 3+3+3): topic overlap 3/4, complementarity 3/3, feasible first test 3/3. A bounded offline comparison is specified; required datasets and domain assumptions must still be checked.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified. Contribution is restricted to the scope stated in the first test.
Connection 3
David Tse
Stanford University
Proposed capability match: Leonid Yavits's computer architecture, content-addressable memory can be paired with David Tse's documented information-theoretic methods, decentralized systems for energy-bounded pathogen surveillance accelerator. The specific contribution is genomics workload correctness; this transfer is an analyst hypothesis.[104][496]
First test and score details
First test
Compare two information-aware sequence-triage baselines before evaluating accelerator speed using public pathogen-read traces and a simulated associative-memory classification kernel. Compare classification disagreement, read coverage and compute per retained sequence with CPU/GPU or conventional memory traces at matched classification fidelity.
Score components
- complementarity
- 3
- feasible first test
- 3
- topic overlap
- 3
Why this rank
Rank 3/10 after semantic revision; analyst score 9 = max(1, 3+3+3): topic overlap 3/4, complementarity 3/3, feasible first test 3/3. A bounded offline comparison is specified; required datasets and domain assumptions must still be checked.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified. His refreshed lab page describes wireless as a previous application and current work on decentralized systems; no active wireless project is assumed. Contribution is restricted to the scope stated in the first test.
Show 7 moreShow fewer external connections
Connection 4
Borivoje Nikolic
University of California, Berkeley
Proposed capability match: Leonid Yavits's computer architecture, content-addressable memory can be paired with Borivoje Nikolic's documented digital integrated circuits, VLSI implementation of communications for energy-bounded pathogen surveillance accelerator. The specific contribution is digital implementation of a selected datapath; this transfer is an analyst hypothesis.[104][436]
First test and score details
First test
Synthesize two small datapath or memory-controller variants under identical constraints using public pathogen-read traces and a simulated associative-memory classification kernel. Compare area, timing slack and switching-activity power estimates with CPU/GPU or conventional memory traces at matched classification fidelity.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 3
Why this rank
Rank 4/10 after semantic revision; analyst score 8 = max(1, 3+3+2): topic overlap 3/4, complementarity 3/3, feasible first test 2/3. The experiment, numerical inputs or identity/scope needs confirmation before execution.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified.
Connection 5
Sophia Shao
University of California, Berkeley
Proposed capability match: Leonid Yavits's computer architecture, content-addressable memory can be paired with Sophia Shao's documented computer architecture, integrated circuits for energy-bounded pathogen surveillance accelerator. The specific contribution is architecture and workload mapping; this transfer is an analyst hypothesis.[104][436]
First test and score details
First test
Map the same kernel to two accelerator organizations without changing numerical precision using public pathogen-read traces and a simulated associative-memory classification kernel. Compare data reuse, memory accesses and latency with CPU/GPU or conventional memory traces at matched classification fidelity.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 3
Why this rank
Rank 5/10 after semantic revision; analyst score 8 = max(1, 3+3+2): topic overlap 3/4, complementarity 3/3, feasible first test 2/3. The experiment, numerical inputs or identity/scope needs confirmation before execution.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified.
Connection 6
Srini Devadas
Massachusetts Institute of Technology
Proposed capability match for Leonid Yavits with Srini Devadas: Secure architecture is a conditional privacy branch for genomic inference; host visibility, unchanged classification behavior and a concrete protected operation must be specified.[104][448]
First test and score details
First test
Specify an untrusted compute host that observes memory-access patterns while classifying synthetic pathogen reads. Compare protected and unprotected versions of the same fixed classifier under identical reads and memory assumptions by classification agreement, traffic and runtime, plus a declared access-pattern leakage proxy. If exact private classification cannot be specified, stop at the threat-model/interface audit.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 6/10 after semantic revision; analyst score 7 = max(1, 2+3+2): topic overlap 2/4, complementarity 3/3, feasible first test 2/3. Secure architecture is a conditional privacy branch for genomic inference; host visibility, unchanged classification behavior and a concrete protected operation must be specified. The experiment, numerical inputs or identity/scope needs confirmation before execution.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified. Post-review scope: Secure architecture is a conditional privacy branch for genomic inference; host visibility, unchanged classification behavior and a concrete protected operation must be specified. This revised proposal awaits independent targeted re-review; simulated outcomes would establish model behavior only, not biological, clinical or deployed benefit.
Connection 7
Pierluigi Nuzzo
University of California, Berkeley
Proposed capability match: Leonid Yavits's computer architecture, content-addressable memory can be paired with Pierluigi Nuzzo's documented cyber-physical systems, design automation for energy-bounded pathogen surveillance accelerator. The specific contribution is design automation and system requirements; this transfer is an analyst hypothesis.[104][436]
First test and score details
First test
Encode latency, energy and failure constraints in a small system model and inject single-component faults using public pathogen-read traces and a simulated associative-memory classification kernel. Compare constraint violations and recovery overhead with CPU/GPU or conventional memory traces at matched classification fidelity.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 7/10 after semantic revision; analyst score 7 = max(1, 2+3+2): topic overlap 2/4, complementarity 3/3, feasible first test 2/3. The experiment, numerical inputs or identity/scope needs confirmation before execution.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified.
Connection 8
Alberto Sangiovanni-Vincentelli
University of California, Berkeley
Proposed capability match: Leonid Yavits's computer architecture, content-addressable memory can be paired with Alberto Sangiovanni-Vincentelli's documented embedded system design, control for energy-bounded pathogen surveillance accelerator. The specific contribution is embedded component partitioning; this transfer is an analyst hypothesis.[104][436]
First test and score details
First test
Partition acquisition, computation and communication across two candidate embedded designs using public pathogen-read traces and a simulated associative-memory classification kernel. Compare interface overhead and end-to-end latency with CPU/GPU or conventional memory traces at matched classification fidelity.
Score components
- complementarity
- 3
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 8/10 after semantic revision; analyst score 7 = max(1, 2+3+2): topic overlap 2/4, complementarity 3/3, feasible first test 2/3. The experiment, numerical inputs or identity/scope needs confirmation before execution.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified.
Connection 9
Jan M. Rabaey
University of California, Berkeley
Proposed capability match for Leonid Yavits with Jan M. Rabaey: Low-power systems expertise is relevant only after declaring an edge deployment and its stages; a genomic-compute trace alone does not establish radio or sensing costs.[4][104][488]
First test and score details
First test
Scope a hypothetical portable sequencing classifier with explicitly separate read arrival, memory/compute and result-radio stages. Use synthetic read arrivals and a fixed classifier cost model; compare two duty-cycle policies at identical read coverage and classification outputs by time-to-result and modelled energy. Sweep sequencing/radio cost assumptions; no existing portable deployment is asserted.
Score components
- complementarity
- 2
- feasible first test
- 2
- topic overlap
- 2
Why this rank
Rank 9/10 after semantic revision; analyst score 6 = max(1, 2+2+2): topic overlap 2/4, complementarity 2/3, feasible first test 2/3. Low-power systems expertise is relevant only after declaring an edge deployment and its stages; a genomic-compute trace alone does not establish radio or sensing costs. The experiment, numerical inputs or identity/scope needs confirmation before execution.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified. Post-review scope: Low-power systems expertise is relevant only after declaring an edge deployment and its stages; a genomic-compute trace alone does not establish radio or sensing costs. This revised proposal awaits independent targeted re-review; simulated outcomes would establish model behavior only, not biological, clinical or deployed benefit.
Connection 10
Ali Javey
University of California, Berkeley
Proposed capability match for Leonid Yavits with Ali Javey: Device expertise is a remote conditional interface contribution; neither a suitable sensing component nor measured device parameters are established by pathogen-read traces.[104][436]
First test and score details
First test
First identify which prospective portable genomic device component would use the candidate's device technology and the electrical interface it must satisfy. For a stipulated sensor/readout component, propagate labelled transfer-curve and noise assumptions into a fixed sequencing-input error model; compare downstream classifier accuracy and modelled power. Reject the branch if no credible component-to-read-error mapping is available.
Score components
- complementarity
- 2
- feasible first test
- 1
- topic overlap
- 1
Why this rank
Rank 10/10 after semantic revision; analyst score 4 = max(1, 1+2+1): topic overlap 1/4, complementarity 2/3, feasible first test 1/3. Device expertise is a remote conditional interface contribution; neither a suitable sensing component nor measured device parameters are established by pathogen-read traces. The experiment, numerical inputs or identity/scope needs confirmation before execution.
Conditions
Proposed fit, not an assertion of a new or active relationship. Independent review pending; forecast confidence low. Partner interest, capacity, data access and any required experimental approvals/resources are unverified. Post-review scope: Device expertise is a remote conditional interface contribution; neither a suitable sensing component nor measured device parameters are established by pathogen-read traces. This revised proposal awaits independent targeted re-review; simulated outcomes would establish model behavior only, not biological, clinical or deployed benefit.
Evidence & open questions
- No current standalone CV was verified.
- Patent-family scope and legal ownership were not investigated.
- The independent reviewer could not reopen the lab overview and publication-list sources. Their original observations remain available, but expertise-claim review is pending and the future-direction claim has partial review coverage.
Inspect claim ratings and independent review
Leonid Yavits works in computer architecture and chip design, including content-addressable and processing-in-memory systems for AI and genomics. [104][105]
Moderate confidenceReview: pendingThe group page does not establish comparative performance of the listed systems.
Review record
- Independent review not recorded for this claim.
An auditable energy-bounded pathogen-surveillance accelerator is a future research hypothesis based on Yavits's and Mutlu's documented methods. [105][107]
Low confidenceReview: partialEnergy, classification fidelity, clinical use and IP constraints remain untested.
Review record
- root: supports. Mutlu’s publication list documents genome-analysis and processing-in-memory systems. The root reviewer could not inspect c-s09 after bounded fetch failures, so the Yavits-specific part lacks independent review coverage.